Recently in Trading Techniques Category

Cameron just left a question for me in the comments to last night's recap. He asked:

I'm looking at your way of thinking about the market. One of the biggest questions I have, though, is regarding your stops.

You said you set your stop at just .20 over your short dollar amount. With all of the intra-day fluctuations, how can you know that it wouldn't have gone up, say, .25 and then nose-dived? It's a perfectly plausible situation.

Said another way, I can place a buy order with a stop, but we easily have 1% swings in a day. If a swing occurs, I have automatically sold for a lower price, only to see the market go back up. Therefore, I'm encouraged to set much broader stops. You set yours for only .20 over the buy price, though, which is why I'm confused.

I often get asked variations of this question, so I thought I'd put my answer in its own post. This was my response:

I couldn't have known that the stock wouldn't have shaken me out. But I also wouldn't have known if I'd have been shaken out if my stop was 50 cents or even a dollar higher. That narrow range candle just provided me with a tight stop. I like tight stops b/c they give me a better opportunity to get high R-multiple winners.

But to your point, how tight of a stop is too tight and will practically guarantee being stopped out? It's hard to answer that question but I generally shy away from any stop that's less than 0.5% of the stock price. I normally like my stops to be about 1% of the stock price. In general, I'm more comfortable with somewhat tighter stops on the QQQQ b/c of how liquid it is and b/c it's less volatile than an "in play" stock would be.

Also, note that I didn't pick a 20 cent stop out of thin air. I went with what the previous candle told me was resistance.

The key thing to remember is that I'm on the hunt for high R-Multiple winners. Because of the way I size my positions (the percent risk model), the tighter my stops the higher the potential reward. In other words if I think a stock has potential to move $2 I'd rather have a 20 cent stop than a $1 stop. Risking 20 cents to make 200 cents ($2) would give me a 10R winner vs only a 2R winner if I had risked $1. (If R was 1% of my equity that would be a 10% gain vs a 2% gain.) For a real example of this see the DGX trades I posted on my Narrow Range Bars post.

So there' a trade-off at work here between stop size and risk/reward multiples. The tighter the stop, the better chance you have of getting a high R winner but you also have a better chance of getting stopped out. Your win ratio is likely to be inversely proportional to how tight your stops are -- looser stops should lead to a higher win rate but also fewer high R-multiple winners.

It's also key that you don't pick your stops willy-nilly. I always use the support/resistance levels that the charts dictate to me.

There's some good advice in Alan Farley's latest article to help "home gamers" (Cramer's phrase) survive the current extra-volatile market. The whole article is worth reading but here are the four bits of advice:

I have four pieces of cautionary advice for my at-home brethren. Listen up, because it's no longer a question of whether you want to take real money out of the market, or just add a few bucks to the weekly shopping budget. These remedial steps must be taken if you want to survive long enough to take advantage of the real opportunity.

1. Wait for the Market Volatility Index (VIX) to drop below 40 and stay there for a week. Massive price swings require equally massive stop losses, which rarely justify the intended positions. Your only alternative is to stand aside and do nothing, no matter how much it hurts to watch others playing those big rallies and selloffs.
2. Forget overnight positions until the index futures stop gapping 2% or more every morning. These price jolts are great news when you're on the right side of the trade, but total devastation if you're on the wrong side. And guess what? You're not smart enough to predict overnight direction from day to day. Neither am I.
3. End your love affair with popular stocks that made you money during the last bull market. In November 2008, these are the issues that will trigger the most painful and unexpected reversals, which happen right after you're absolutely convinced your position is the right play. The bottom line: They see you coming, sucker.
4. Get control of the time element in your market strategy. You're getting killed because you have no patience and forgot how to sit on your hands when your trading edge isn't in play. Realistically, it could be months before the market works for you again. Would you rather wait it out and survive, or stay busy and get crushed?

My thoughts:

1. I like that part about watching the VIX. A few weeks ago I decided to cut my risk per trade (R) in half (from 1% to 0.5% of my equity) until the VIX breaks that top it's trying to put in. That means a close below 44. (Dr. Brett recently wrote about when traders should cut their risk.)
2. Tough to argue with what he said about overnight holds. This is certainly a market for intraday trades.
3. Reminds me of what I said about AAPL and GOOG back in July. Add DRYS and its ilk to the list too. It's better to just have an open mind and grab on to whichever stocks become the new leaders.
4. Yep, preservation of capital must be the top priority. If you can't sit out at least lower your risk.

I was recently asked about how I determine my "universe" of stocks. That is the stocks that I consider fair game to be traded. I use different filters for swing trading and day trading but they do have similar criteria. In both cases I'm trying to find liquid, volatile stocks. Once the filter's applied I then look for certain patterns / setups to trade.

For swing trade candidates I use the following criteria in TeleChart:

• Price Per Share: about $5 to Max
• 5 Day Average Volume: 75th percentile & higher (right now that works out to about 466,000 shares per day)
• Price Volatility: 25th percentile & higher
• Volume (Dollars) 5-Day: 75th percentile & higher

Here's a shot of the filter. The count column shows how many stocks are left after applying each row in the filter.

That typically narrows the market down to about 1,000 stocks. Then I look for setups among that set of stocks. I gave a lot of details about the setups I scan for back in the day.

For day trading I'm trying to find liquid stocks which are abnormally active and volatile. To do that I use the following two scans in Trade-Ideas:

Bearish Scanner's Filter:

• Min Price: $10
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: -0.5% (negative means down at least 0.5%)

Bullish Scanner's Filter:

• Min Price: $5
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: 1%

Sometimes I'll tweak these settings to give more or fewer alerts. I usually do that by adjusting the "Min Current Volume". Details about the actual alerts & setups I track for daytrading are on my "how I work" page. If you use Trade-Ideas you can paste the following link into the "Collaborate" feature and load that filter and all the alerts I track:

Bearish -- grab the URL of this link

Bullish -- grab the URL of this link

One thing I have always preached on the blog, less so on Twitter, where I bang out more trading ideas and market thoughts is that shorting stocks is hard. I think it’s harder than any aspect of learning the market. It’s dangerous. Let this morning be the only reminder you should EVER need.

Howard's right about shorting being difficult. I think it's so hard because it's not simply the opposite of going long. (I won't even go into the whole thing about your losses while short being theoretically infinite. Been there... done that. Use a stop to limit your losses, use proper position sizing, ONLY short LIQUID stocks and you'll be fine.) What makes shorting tricky is that bear moves often have violent (short-covering) rallies because the psychology of the crowd trading a down market is different than that of a bull market. You have to be quick on your feet when shorting. My motto is "stick & move".

Many traders love to buy breakouts in bull markets. (Whether that's actually a good strategy is debatable). In my experience swing trading, the opposite of that strategy, shorting breakdowns through resistance, will often lead you right into a snapback rally and, as MaoXian used to say "the quickest loss ever". That's why I often make note of all the people who are initiating shorts after the market has already fallen to a major support level. We saw that this week as I noted in some of the morning watchlists.

My contention is that if you were caught short Friday morning you should consider your losses as tuition paid to the school of hard knocks. Learn from that expensive lesson, take your losses and hopefully survive to trade another day. I'll stop short of saying that the losses were deserved but there were plenty of warnings to at least cover your shorts if not to get long. There are so many good sources of market information these days, both on the web and, yes, even on TV. I'm not saying to blindly follow somebody else's opinion but it can be helpful to see what others are doing based on what they see. Here are just a few of the recent warning signs:

• T2108 dropping below 20 -- Ah, good old reliable T2108. I've been watching it closely as we've sold off. On Wednesday I noted that it finally hit the point where wise shorts would want to cover. It pays to find a good overbought/oversold indicator and heed its warnings. (You may need different indicators or settings for different timeframes.) Sell at overbought and cover at oversold. A couple of years ago I decided to force myself to put some IRA money to work every time T2108 broke 20. It hasn't failed me yet.
• The Fed (Plunge Protection Team) has interfered announced stimulus packages around the July lows a couple times. On Thursday and earlier in the week there was talk of more PPT action.
• The PPT took action earlier in the week and last week. Just look at the orchestration of the LEH and AIG situations, etc.
• Extreme Volatility -- The moves all week were nothing short of violent. Positions were whipsawed all over the place. That in & of itself would be reason enough to lighten up on positions if not move to the sidelines. That kind of volatility is often a sign of a trend reversal, not of a continuation of the previous trend. That's why so many traders watch the VIX. Tons of people noted the spike in the VIX on Thursday.
• Corey from the 'Afraid to Trade' blog warned 'Use Extreme Caution in the Week Ahead.' He gave many good reasons, including the Federal Reserve interest rate decision, the quadruple witching options expiration and headline risk from troubled financial firms. His crystal ball was working well when he wrote "We could see a week ahead that will be discussed years later - as such, if you are a newer trader, it might be best to switch to simulation mode this week or use this week as a training experience, rather than risking real capital in an environment that could swing violently up and down due to market events scheduled to happen this week."
• Bullish Technical Divergences -- Dr. Brett pointed out some bullish divergences he was seeing in the market. Perhaps most important were what he called 'those fuzzy indicators' -- "Traffic on the blog is way up, reflecting trader uncertainty and desire for information. I just fielded my fourth media interview request in two days. During quiet and bullish market periods, I don't get four requests in a month." On Tuesday morning I also noted my blog's traffic spiked as did Barry Ritholtz. I've often joked about making some kind of sentiment indicator based on my site's traffic ebb & flow & referral logs. It's not a bad idea and I think it would be especially useful if it were based on a major financial site's traffic data.
• Dennis Gartman telling folks to "be small" -- Gartman was on CNBC's 'Fast Money' early in the week saying that he was scared of the market's movement and he was "being small" and planned to "get smaller". His advice to others was to "be small" in this market.
• Jeff Macke, also on 'Fast Money' was warning people not to play (trade) if they didn't understand the (changing) rules of the game. He was referring to all the headline risk from PPT action and the volatility caused by rumors -- many of which were spread by CNBC during the trading sessions. Ironically, Macke was kicking himself Thursday night for not following his own advice and getting caught short.

I fully believe that had the PPT not acted on Thursday night the market was *destined* to move higher in the short term on its own. Still being short on expiration Friday in this environment was just asking for trouble. So what's a trader to do? Like I said earlier, stick & move. I think it makes much more sense to short bounces back to a trendline or moving average. William O'Neil's book on short selling talks about using retracements to the 50 and/or 200-day moving average as a place to initiate shorts. By the time the stock (or whatever instrument you're trading) is extended from its moving average it's time to cover. Then you can wait for another bounce to reload.

The strategy to cover & reload makes more sense when you're short due to the fact that if you're dead right with your call on the stock dropping the most you can gain is 100%. The odds of that happening are slim and if it ever did happen you'd likely have to ride out some severe short squeezes. But you might be able to stick & move your way to more that a 100% gain by reloading multiple times. In theory, you could catch 15 10% moves lower in a stock that's falling, retracing & falling anew.

Short sellers need to be nimble, pick their entry and exit spots wisely and heed signs of impending reversals. Trying to short breakdowns of an already extended market is a sucker's play -- as is overstaying your welcome while short.

I got this question the other day:

When you short a stock at 30.00 and you want to place a stop limit in case it jumps up past your max loss of say, 31.00, how exactly is that order entered (everytime I try to do this, it sells (Ed: he means buys) my stock immediately - I'm doing something really wrong). Do I put a limit price of 31.00 and a stop price above that, i.e. 31.50 (thus saying it's ok to repurchase any shares between 31.00 and 31.50)?

This is one of the trickier orders to place. In that situation you would want to enter a buy stop limit order (buy on stop with a limit). The stop would be at 31.00 and the limit would be 31.50. Here's a screen shot of a similar order to cover QQQQ with a buy stop @ 47 with a limit of 47.10:

What would happen here is that once the stop price of 47.00 is hit a buy limit order of 47.10 would be entered. As with any limit order, you're guaranteed price but not execution, so you may not be filled. If you wanted to be sure that you got out (covered) just use a buy stop of $47.00.

In case you missed it in the links section the other day, be sure to read the interview with Michelle B over on the StockTickr blog.

The other day Downtown Trader asked me how I handled trailing my stop on that DGX short. I pretty much followed exactly what's in my post about how I take partial profits. I moved the stop to breakeven once my gain was equal to my initial risk of 50 cents per share. As it dropped further I moved the stop to lock in one-third of my profit. Why one-third? It's a value I came up with which I think allows a stock to have a normal retracement without stopping me out. It's sort of an arbitrary level, although I got the idea from the Fibonacci levels. I've found that method to work reasonably well after much trial and error with trailing stops. I read somewhere that you can consider a retracement to be normal as long as it didn't break the 62% Fib level. So I figured that I could trail my stops somewhere around that level to allow stocks to have their normal fluctuations. I generally get more aggressive with the trailing stops as the day wears on. I'll move them to 50% by 3:00 and just keep pushing them until the close -- assuming I don't bail before then like I did with DGX.

The other day I got the following email about narrow range bars:

I was going through your site and others and I keep on hearing, no entry on narrow range bars. What is a narrow range bar, and could you send me a picture of one?

A narrow range bar (candle) is simply a bar which has a range from high to low that's much less than the average bar for a given equity. You'll often see people, like Dr. Sen, tracking NR7 bars. NR7 means the narrowest range of the last seven bars. For example, here are some NR7s from last Friday:

So now that you know what narrow range candles are, why should you care about them? Since stocks generally cycle between periods of low volatility and high volatility it can be beneficial to identify when the high volatility times are approaching. Often times, especially for trending charts, narrow range periods -- periods of range contraction -- will presage range expansion. You can think of a stocks movement like a spring / slinky -- it need to recoil (contract) in order to build up enough potential energy for its next expansion. However, it's important thing to note that range contraction doesn't tell you the direction of the impending expansion only that an expansion may be coming. It's up to the trader to use other means (indicators, common sense?) to catch the expansion in the right direction.

One of the nice things about trading against narrow range bars is that they give you a tight stop. If you size your positions like I do, the smaller the stop means you can buy (or short) more shares. It's the trades with the really tight stops that let you make those huge R-Multiple trades that often make or break a trader. Here's an example from a trade I took last week.

I'm going to use the actual entries taken by myself and by Butterboy but instead of using my actual exits, I'll assume I held until the close. So let's say my equity is $100,000 and my R, the amount I want to risk per trade, is 1% or $1,000. Given that, here's the first trade, based off of my actual entry parameters:

• Entry at 53.96
• Stop was 50 cents higher so I could short 1000/.50 or 2,000 shares
• DGX closed at 49.64, which is $4.32 lower. So that trade returned $8,640 or 8.64 times my risk (8.64R)

Here's the DGX trade using Butterboy's entry criteria. He used a 15 minute chart and entered lower than I did but his stop was also much tighter:

• Entry at 53.40
• I believe his stop was 21 cents higher, just above the 11:30 bar which he entered against. Given that one could short 1,000/.21 or just over 4,700 shares.
• DGX closed at $3.76 beneath the entry so that trade returned $17,672 or 17.67 times the initial risk (17.64R).

It should be clear why so many of us like narrow range bars and why I scan for them. That second trade had a worse (lower) entry but because of its tighter stop it was almost twice as profitable. It also shows why I love the percent risk position sizing model.

Here's an article on range contraction written by Oswald S. Castillo, TheStockStalker and very possibly the #1 fan of range contraction trading. Oswald also has a webinar on his site entitled "Principles of Trading Range Contraction". It's a good video but you have to register to view it.

You may also find Alan Farley's work on "coiled springs" and NR7s helpful. He goes in-depth on coiled springs in his book, The Master Swing Trader.

P.S. I almost forgot to answer the question of: How Narrow is Narrow?

For daytrading I like to find candles that are 1% or less than the price of the stock. So for a $20 stock any candles with a range of 20 cents or less really get my attention. In cases like those I just need small moves in the stock to rack up some nice gains. I'll typically pass on candles with a range much greater than 1% because they'll require a much larger move in the stock in order for me to get those large R-multiples.

For swing trading (looking at daily charts) I don't have a concrete answer on how narrow is narrow. But you could probably come up with some value based on the stock's average true range (ATR) or some other method. For example, you could define narrow as less than half a stocks ATR.

NRx (narrowest range on the last X candles) is always a good way to go as well, and it's easy to scan for.

Michelle B submits: I remember reading about a French housewife who when asked by her husband every morning what they would be having for dinner would always give the same tireless reply: "It depends on what jumps into my market basket." Some traders prefer to trade from a prepared watchlist. For me, trading is interesting because I do not know what will jump into my market basket. My preparation is my confidence in my ability to execute perceived opportunities according to my risk parameters.

Using the premarket top gainers/losers scan, I search the NASDAQ, NYSE, and AMEX market stalls for any tasty morsels and look for any stocks moving on the highs/lows scan. I also check Briefing.com and MarketWatch Newsfinder for news that has happened since the close of the previous trading day---identifying which stocks the market considers to be newsworthy is one of the many trading skills which can be developed.

In addition, I note via Briefing.com calendars such events as economic reports, earnings, Fedspeak, splits, upgrades/downgrades, and conferences. Usually, I have around ten to twelve trading candidates. Sometimes, there are none or a very small number. In that case, I patiently wait until the market has a new batch of fresh candidates---I will not make do with stale merchandise.

I list the stock symbols of my candidates and the important economic events for that day in my journal. Next to each stock symbol, I put the results of my basic Yahoo financial research---reason for price movement, float size, and short interest. I identify support and resistance on various timeframes. Sometimes, I check Yahoo Finance message boards and blogs to find out why a stock may be moving. Though in that case, I am always on my guard to weed the chaff from the wheat.

Then, I place my candidates on the chart pages of my Advanced Trading Platform (ATP). Four thirty-minute charts go on pages two through four. The main page contains the following windows: order entry, order status, real time portfolio, high/low scan, top gainers/losers scan, watchlist containing, in addition to my trading candidates, the major indices and technical indicators like VIX and TICK. Last but not least, is a chart window for the most promising candidate.

Discovering that ININ---set to gap up because it has given higher earnings' guidance---has a very small float and moderately high short interest, coupled with my knowledge that such stocks often move very strongly on low volume days, like Fridays, pre-holidays, and summer days, makes ININ the star attraction at this stage.

The market opens, and since I am particularly interested in seeing if ININ can trigger a trade, it gets top attention with a one-minute chart (above) on my main trading page. Meanwhile, I am flipping through my chart pages to see if anything is close to triggering---nothing yet. I check the one-minute chart of ININ, and I see the choppy pennant being formed at 10 A.M. The pennant range is around .25. That's enough of a sign for me to buy half of my lot as close to the bottom of this formation at 13.65. Buying a partial lot before the actual breakout happens makes sense in high-probability trades, especially if the desired lot size is substantial. I am set to buy the rest via an automatic buy stop if it can clear the top of the pennant at 13.87 on the 5 minute (see below). Trades moving strongly in the morning require monitoring via shorter time frames.

Fifteen minutes later, it does, and my full lot is in, average cost around 13.78. Just before the breakout, the pennant was around seventy-five percent completed and the volume dried up. Triangles/pennants need to be not completely formed or else they usually just fizzle out and do not support strong breakouts. The pause in volume is what my trading buddy calls a lullipop, a lull before the breakout, where the equilibrium between buyers and sellers end and the buyers take over.

Part of my prepurchase research of ININ was its support and resistance based on weekly price level resistance. Since the open of the July 24, 2006 weekly, bearish, high-volume candle was 14.81 (see below), therefore trapping lots of traders at that level, I determined the resistance to be around 14.80. The target of 14.80 also matches the length of the pole on which the flag formed on Friday's intraday charts, forming a measured move and confirming the resistance.

As the ININ trade continues, I am monitoring the price action on different timeframes, and have my mental stop under the flag for the first half lot at 13.53 and then just above the flag at 13.89 for the complete lot. Price is smoothly rising---each candle opens near but not with much overlapping to the close of the previous candle. I put my sell limit in for the full lot just under the target price. My offer gets lifted fairly easily, and I am out with a $1.00 move gotten in about 40 minutes. Since the risk was .25 for the complete lot, it was a +4R trade. For quickly moving stocks, I will often forgo hard stops, but it is a risk because my net connection could be interrupted. Then my carefully constructed trade will exist only in cyberspace floating without my pilotage. So, I am working on the artful science of consistently setting hard regular/trailing stops with upper triggers via bracketed orders to take care of that possibility.

My market basket is now empty until the next fresh thing jumps into it.

There seems to be a lot of controversial over the concept of R-Multiples. I've been seeing people complain about them for months now and I've been meaning to write a post about "R". I really wanted to do it last week but I'm glad I didn't get to it because this week a raging debate about R has popped up. Glenn, at DehTrader can serve as the poster child for the anti-R crew. Here's part of his recent rant against R-Multiples (emphasis is mine):

I post real numbers as opposed to R values, I always have. I like real numbers, I understand real numbers and I see truth in real numbers and I think the reader does too. As a reader of many blogs I find zero value in any post or summaries containing R values, I don't see any point in sharing that information. I suppose if I posted in R values I could look like a pretty good trader, but we all know I am a struggling trader. R can mean anything so why even bother with it... R stands for bullsh!t imo and that's my rant (that and ads haha). The best blogs out there post real numbers, Boogtser, JC (NYSE), the Kirkster all come to mind.

He's joined by folks like Paul who left this comment over on Ugly's post about R multiples:

I believe dollar values are more important than R value. I agree that the actual $ value is meaningless. However, R values are subjective and don’t give you a true idea on how successful the trade was. If you defined your risk at 15 cents and made 30 cents on the trade, while another person made 50 cents but decided his risk would be 50 cents, R values would say the guy who made 30 cents was more successful. I have a problem with that. It could very well be that the guy who only risked 15 cents is playing it too safe and his 2R gain was a bad trade.

So that gives you an idea of the anti-R sentiment. I'm going to explain why I think R-Multiples are so useful and why I use them in my trading and on this site.

What is R?

R is simply the dollar risk per trade. It's nothing but a reward-to-risk ratio. I first heard it called "R" in Van Tharp's book "Trade Your Way to Financial Freedom". In another of his books, "Financial Freedom Through Electronic Day Trading", Dr. Tharp reveals the great secret of trading:

The golden rule of trading is to keep losses at a level of 1 R as often as possible and to make profits that are high-R multiples.

You often hear (read) that traders should only look for trades with a reward/risk ratio of at least 2 or 3 to 1. Expressing your results in terms of how many times your risk allows you to easily see how well your trades measure up to such a standard. So when I look at my results in terms of multiples of R I can easily tell how good or bad the trades were. I like to think of R-Multiples as telling you the efficiency of your system.

So why not just use dollars?

Expressing my results in dollars would achieve the same result if I always risked the same amount of money. But what if I triple my account and therefore trade larger positions compared to when I started trading? Or what if I hit a rough spot and decide to cut my share size down while I ride out the storm? Then the dollar results won't easily tell me how trades from one period of time compared to another period of time. But if I use R making such comparisons is simple. Either my trades passed the risk / reward ratio test or they didn't. The actual number of dollars at risk doesn't matter, how many multiples of the dollars at risk does.

Along the same lines, recording trades in terms of R-Multiples allows you to easily calculate your system's expectancy. (Follow the link for why you should care about expectancy.)

Also, as Rx said:

talking and thinking in terms of R-multiple when you discuss about profits is an excellent approach - that by itself makes you focus on risk and money management - the actual "grail" to successful trading.

That is a very important point. Whenever I see people posting dollar returns, especially losses, that are all over the place the first thing I ask myself is "I wonder what his risk per trade is". It's almost a certainty that those traders aren't focusing on risk and as a result keep having huge losses. The mere fact that you have to define R and then place a stop to keep your loss to 1R is probably too constraining for those gamblers traders. Dr. Tharp says about determining your initial stop-loss point as soon as you enter a trade, which, by definition woud give you a 1 R loss:

This principle is so important that if you cannot follow it, then you might as well give up the idea of electronic day trading right away.

The reason I use R on the blog is because I don't want to discuss dollars or my account size on the site (as Ugly stated). That's nobody's business but mine. Also, it makes it easy for people to figure out what they could have made or lost on a trade with their own account size and risk per trade amount. If you see a trade that returned 3R all one has to do is plug in their dollar risk per trade to figure out what they could have made / lost.

To the R Haters

Let me address the "alleged" issues which I quoted above...

Glenn thinks that R is just some made up number and could mean anything. He likes "real" numbers. While it may be fun to see that somebody made $10,000 on a trade that in and of itself doesn't tell you how good that trade was. What if that person risked $30,000 to make that $10,000? Or what if they risked $1,000 to make that $10,000? Those are two very different trades. Sure they both made the same amount of money but isn't the second trade a much more efficient use of capital?

What if somebody is trading $500,000 lots to make $1,000 in profit? It may be nice to see somebody saying that they made $1,000 here and $1,000 there but damn(!) that's an inefficient use of capital. So while R could mean anything in terms of dollars, in my humble opinion what really matters is how many multiples of R were made or lost. That tells you the quality of a trade or system.

Glenn also states that if he reported his trades in terms of R he could appear to be a good trader. I'm sorry to tell him that's simply not the case. If you lost money that means your expectancy, which is just your average return expressed in R-Multiples, was negative.

Paul said that "R values are subjective and don’t give you a true idea on how successful the trade was". That is exactly wrong. R-multiples are the very thing that tells you exactly how successful a given trade was, if you choose to grade on a risk/reward basis.

Percentages vs. Dollars

This debate about R reminds me of a conversation I had a couple of weeks ago. I was in a presentation for Trade-Ideas' new tool, the Odds Maker. They were showing how you could backtest all these different scenarios with the tool. The results were expressed in average dollars won or lost. Another viewer and I asked about seeing the results in percentages. They kept saying that perhaps they would do that in a later revision. I kept harping on it because to me seeing the results in dollars was of little use for the way I size my positions

The argument from the presenter was that all you had to do was multiply the average dollar return by your average lot size to figure out how much money you could have made with a given system you were testing. I had to disagree because my lot size can vary drastically depending on how far away my stop loss is. Here's a situation which could be problematic -- I trade Google with a 2 point stop (which is only about half of a percent) and get lucky and make 6 points of profit. All of my other trades are on stocks under $50 with stops less than 50 cents. I could have some combination of winners and losers mixed in there... most of them probably well less than $6. That $6 gain may skew the results when presented as average dollars won. That's an over-simplification and there are all kinds of possible permutations. But I hope my point is clear that looking at the results in terms of average dollars won/lost may not tell accurately tell you the story.

So how can we make the results clearer? Simple, express them in percentages. That way, regardless of how many shares were traded or the prices of the stocks traded the results can be equalized across all the trades. I feel much better being able to say , "OK, this system would have returned X%" instead of "X number of points.

We debated the merits of each way of reporting for a few minutes and at one point somebody said, well , for this release we're aiming for the "lowest common denominator". In other words, the average person can't think in percentages, so we're just gonna report in points. I was like, F the average person, make it work the "right" way! The funny thing is that after debating all of that the software actually could express the results in percentage terms. We just had to switch a setting.

So my point of that little story is that I always prefer to think in terms of percentages in stead of points. I always see people talking about number of shares of point moves. For example, you might hear somebody exclaim "Google is up 5 points!!!" I don't see that as anything to get excited about. That just over a 1% move -- a normal fluctuation. You'll hear similar things from reporters talking excitedly about the Dow being up some triple-digit amount. The Nasdaq may actually be up a lot more on a percentage basis but they'll just say, eh, the Nasdaq is "just" up 30 points.

Looking at the percentages makes those kind of comparisons easier. R-Multiples do the same thing for traders. They can accurately compare their own trades and they can take another trader's results expressed in R and easily relate them to their own system.

Cameron just left a question for me in the comments to last night's recap. He asked:

I'm looking at your way of thinking about the market. One of the biggest questions I have, though, is regarding your stops.

You said you set your stop at just .20 over your short dollar amount. With all of the intra-day fluctuations, how can you know that it wouldn't have gone up, say, .25 and then nose-dived? It's a perfectly plausible situation.

Said another way, I can place a buy order with a stop, but we easily have 1% swings in a day. If a swing occurs, I have automatically sold for a lower price, only to see the market go back up. Therefore, I'm encouraged to set much broader stops. You set yours for only .20 over the buy price, though, which is why I'm confused.

I often get asked variations of this question, so I thought I'd put my answer in its own post. This was my response:

I couldn't have known that the stock wouldn't have shaken me out. But I also wouldn't have known if I'd have been shaken out if my stop was 50 cents or even a dollar higher. That narrow range candle just provided me with a tight stop. I like tight stops b/c they give me a better opportunity to get high R-multiple winners.

But to your point, how tight of a stop is too tight and will practically guarantee being stopped out? It's hard to answer that question but I generally shy away from any stop that's less than 0.5% of the stock price. I normally like my stops to be about 1% of the stock price. In general, I'm more comfortable with somewhat tighter stops on the QQQQ b/c of how liquid it is and b/c it's less volatile than an "in play" stock would be.

Also, note that I didn't pick a 20 cent stop out of thin air. I went with what the previous candle told me was resistance.

The key thing to remember is that I'm on the hunt for high R-Multiple winners. Because of the way I size my positions (the percent risk model), the tighter my stops the higher the potential reward. In other words if I think a stock has potential to move $2 I'd rather have a 20 cent stop than a $1 stop. Risking 20 cents to make 200 cents ($2) would give me a 10R winner vs only a 2R winner if I had risked $1. (If R was 1% of my equity that would be a 10% gain vs a 2% gain.) For a real example of this see the DGX trades I posted on my Narrow Range Bars post.

So there' a trade-off at work here between stop size and risk/reward multiples. The tighter the stop, the better chance you have of getting a high R winner but you also have a better chance of getting stopped out. Your win ratio is likely to be inversely proportional to how tight your stops are -- looser stops should lead to a higher win rate but also fewer high R-multiple winners.

It's also key that you don't pick your stops willy-nilly. I always use the support/resistance levels that the charts dictate to me.

There's some good advice in Alan Farley's latest article to help "home gamers" (Cramer's phrase) survive the current extra-volatile market. The whole article is worth reading but here are the four bits of advice:

I have four pieces of cautionary advice for my at-home brethren. Listen up, because it's no longer a question of whether you want to take real money out of the market, or just add a few bucks to the weekly shopping budget. These remedial steps must be taken if you want to survive long enough to take advantage of the real opportunity.

1. Wait for the Market Volatility Index (VIX) to drop below 40 and stay there for a week. Massive price swings require equally massive stop losses, which rarely justify the intended positions. Your only alternative is to stand aside and do nothing, no matter how much it hurts to watch others playing those big rallies and selloffs.
2. Forget overnight positions until the index futures stop gapping 2% or more every morning. These price jolts are great news when you're on the right side of the trade, but total devastation if you're on the wrong side. And guess what? You're not smart enough to predict overnight direction from day to day. Neither am I.
3. End your love affair with popular stocks that made you money during the last bull market. In November 2008, these are the issues that will trigger the most painful and unexpected reversals, which happen right after you're absolutely convinced your position is the right play. The bottom line: They see you coming, sucker.
4. Get control of the time element in your market strategy. You're getting killed because you have no patience and forgot how to sit on your hands when your trading edge isn't in play. Realistically, it could be months before the market works for you again. Would you rather wait it out and survive, or stay busy and get crushed?

My thoughts:

1. I like that part about watching the VIX. A few weeks ago I decided to cut my risk per trade (R) in half (from 1% to 0.5% of my equity) until the VIX breaks that top it's trying to put in. That means a close below 44. (Dr. Brett recently wrote about when traders should cut their risk.)
2. Tough to argue with what he said about overnight holds. This is certainly a market for intraday trades.
3. Reminds me of what I said about AAPL and GOOG back in July. Add DRYS and its ilk to the list too. It's better to just have an open mind and grab on to whichever stocks become the new leaders.
4. Yep, preservation of capital must be the top priority. If you can't sit out at least lower your risk.

I was recently asked about how I determine my "universe" of stocks. That is the stocks that I consider fair game to be traded. I use different filters for swing trading and day trading but they do have similar criteria. In both cases I'm trying to find liquid, volatile stocks. Once the filter's applied I then look for certain patterns / setups to trade.

For swing trade candidates I use the following criteria in TeleChart:

• Price Per Share: about $5 to Max
• 5 Day Average Volume: 75th percentile & higher (right now that works out to about 466,000 shares per day)
• Price Volatility: 25th percentile & higher
• Volume (Dollars) 5-Day: 75th percentile & higher

Here's a shot of the filter. The count column shows how many stocks are left after applying each row in the filter.

That typically narrows the market down to about 1,000 stocks. Then I look for setups among that set of stocks. I gave a lot of details about the setups I scan for back in the day.

For day trading I'm trying to find liquid stocks which are abnormally active and volatile. To do that I use the following two scans in Trade-Ideas:

Bearish Scanner's Filter:

• Min Price: $10
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: -0.5% (negative means down at least 0.5%)

Bullish Scanner's Filter:

• Min Price: $5
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: 1%

Sometimes I'll tweak these settings to give more or fewer alerts. I usually do that by adjusting the "Min Current Volume". Details about the actual alerts & setups I track for daytrading are on my "how I work" page. If you use Trade-Ideas you can paste the following link into the "Collaborate" feature and load that filter and all the alerts I track:

Bearish -- grab the URL of this link

Bullish -- grab the URL of this link

One thing I have always preached on the blog, less so on Twitter, where I bang out more trading ideas and market thoughts is that shorting stocks is hard. I think it’s harder than any aspect of learning the market. It’s dangerous. Let this morning be the only reminder you should EVER need.

Howard's right about shorting being difficult. I think it's so hard because it's not simply the opposite of going long. (I won't even go into the whole thing about your losses while short being theoretically infinite. Been there... done that. Use a stop to limit your losses, use proper position sizing, ONLY short LIQUID stocks and you'll be fine.) What makes shorting tricky is that bear moves often have violent (short-covering) rallies because the psychology of the crowd trading a down market is different than that of a bull market. You have to be quick on your feet when shorting. My motto is "stick & move".

Many traders love to buy breakouts in bull markets. (Whether that's actually a good strategy is debatable). In my experience swing trading, the opposite of that strategy, shorting breakdowns through resistance, will often lead you right into a snapback rally and, as MaoXian used to say "the quickest loss ever". That's why I often make note of all the people who are initiating shorts after the market has already fallen to a major support level. We saw that this week as I noted in some of the morning watchlists.

My contention is that if you were caught short Friday morning you should consider your losses as tuition paid to the school of hard knocks. Learn from that expensive lesson, take your losses and hopefully survive to trade another day. I'll stop short of saying that the losses were deserved but there were plenty of warnings to at least cover your shorts if not to get long. There are so many good sources of market information these days, both on the web and, yes, even on TV. I'm not saying to blindly follow somebody else's opinion but it can be helpful to see what others are doing based on what they see. Here are just a few of the recent warning signs:

• T2108 dropping below 20 -- Ah, good old reliable T2108. I've been watching it closely as we've sold off. On Wednesday I noted that it finally hit the point where wise shorts would want to cover. It pays to find a good overbought/oversold indicator and heed its warnings. (You may need different indicators or settings for different timeframes.) Sell at overbought and cover at oversold. A couple of years ago I decided to force myself to put some IRA money to work every time T2108 broke 20. It hasn't failed me yet.
• The Fed (Plunge Protection Team) has interfered announced stimulus packages around the July lows a couple times. On Thursday and earlier in the week there was talk of more PPT action.
• The PPT took action earlier in the week and last week. Just look at the orchestration of the LEH and AIG situations, etc.
• Extreme Volatility -- The moves all week were nothing short of violent. Positions were whipsawed all over the place. That in & of itself would be reason enough to lighten up on positions if not move to the sidelines. That kind of volatility is often a sign of a trend reversal, not of a continuation of the previous trend. That's why so many traders watch the VIX. Tons of people noted the spike in the VIX on Thursday.
• Corey from the 'Afraid to Trade' blog warned 'Use Extreme Caution in the Week Ahead.' He gave many good reasons, including the Federal Reserve interest rate decision, the quadruple witching options expiration and headline risk from troubled financial firms. His crystal ball was working well when he wrote "We could see a week ahead that will be discussed years later - as such, if you are a newer trader, it might be best to switch to simulation mode this week or use this week as a training experience, rather than risking real capital in an environment that could swing violently up and down due to market events scheduled to happen this week."
• Bullish Technical Divergences -- Dr. Brett pointed out some bullish divergences he was seeing in the market. Perhaps most important were what he called 'those fuzzy indicators' -- "Traffic on the blog is way up, reflecting trader uncertainty and desire for information. I just fielded my fourth media interview request in two days. During quiet and bullish market periods, I don't get four requests in a month." On Tuesday morning I also noted my blog's traffic spiked as did Barry Ritholtz. I've often joked about making some kind of sentiment indicator based on my site's traffic ebb & flow & referral logs. It's not a bad idea and I think it would be especially useful if it were based on a major financial site's traffic data.
• Dennis Gartman telling folks to "be small" -- Gartman was on CNBC's 'Fast Money' early in the week saying that he was scared of the market's movement and he was "being small" and planned to "get smaller". His advice to others was to "be small" in this market.
• Jeff Macke, also on 'Fast Money' was warning people not to play (trade) if they didn't understand the (changing) rules of the game. He was referring to all the headline risk from PPT action and the volatility caused by rumors -- many of which were spread by CNBC during the trading sessions. Ironically, Macke was kicking himself Thursday night for not following his own advice and getting caught short.

I fully believe that had the PPT not acted on Thursday night the market was *destined* to move higher in the short term on its own. Still being short on expiration Friday in this environment was just asking for trouble. So what's a trader to do? Like I said earlier, stick & move. I think it makes much more sense to short bounces back to a trendline or moving average. William O'Neil's book on short selling talks about using retracements to the 50 and/or 200-day moving average as a place to initiate shorts. By the time the stock (or whatever instrument you're trading) is extended from its moving average it's time to cover. Then you can wait for another bounce to reload.

The strategy to cover & reload makes more sense when you're short due to the fact that if you're dead right with your call on the stock dropping the most you can gain is 100%. The odds of that happening are slim and if it ever did happen you'd likely have to ride out some severe short squeezes. But you might be able to stick & move your way to more that a 100% gain by reloading multiple times. In theory, you could catch 15 10% moves lower in a stock that's falling, retracing & falling anew.

Short sellers need to be nimble, pick their entry and exit spots wisely and heed signs of impending reversals. Trying to short breakdowns of an already extended market is a sucker's play -- as is overstaying your welcome while short.

I got this question the other day:

When you short a stock at 30.00 and you want to place a stop limit in case it jumps up past your max loss of say, 31.00, how exactly is that order entered (everytime I try to do this, it sells (Ed: he means buys) my stock immediately - I'm doing something really wrong). Do I put a limit price of 31.00 and a stop price above that, i.e. 31.50 (thus saying it's ok to repurchase any shares between 31.00 and 31.50)?

This is one of the trickier orders to place. In that situation you would want to enter a buy stop limit order (buy on stop with a limit). The stop would be at 31.00 and the limit would be 31.50. Here's a screen shot of a similar order to cover QQQQ with a buy stop @ 47 with a limit of 47.10:

What would happen here is that once the stop price of 47.00 is hit a buy limit order of 47.10 would be entered. As with any limit order, you're guaranteed price but not execution, so you may not be filled. If you wanted to be sure that you got out (covered) just use a buy stop of $47.00.

In case you missed it in the links section the other day, be sure to read the interview with Michelle B over on the StockTickr blog.

The other day Downtown Trader asked me how I handled trailing my stop on that DGX short. I pretty much followed exactly what's in my post about how I take partial profits. I moved the stop to breakeven once my gain was equal to my initial risk of 50 cents per share. As it dropped further I moved the stop to lock in one-third of my profit. Why one-third? It's a value I came up with which I think allows a stock to have a normal retracement without stopping me out. It's sort of an arbitrary level, although I got the idea from the Fibonacci levels. I've found that method to work reasonably well after much trial and error with trailing stops. I read somewhere that you can consider a retracement to be normal as long as it didn't break the 62% Fib level. So I figured that I could trail my stops somewhere around that level to allow stocks to have their normal fluctuations. I generally get more aggressive with the trailing stops as the day wears on. I'll move them to 50% by 3:00 and just keep pushing them until the close -- assuming I don't bail before then like I did with DGX.

The other day I got the following email about narrow range bars:

I was going through your site and others and I keep on hearing, no entry on narrow range bars. What is a narrow range bar, and could you send me a picture of one?

A narrow range bar (candle) is simply a bar which has a range from high to low that's much less than the average bar for a given equity. You'll often see people, like Dr. Sen, tracking NR7 bars. NR7 means the narrowest range of the last seven bars. For example, here are some NR7s from last Friday:

So now that you know what narrow range candles are, why should you care about them? Since stocks generally cycle between periods of low volatility and high volatility it can be beneficial to identify when the high volatility times are approaching. Often times, especially for trending charts, narrow range periods -- periods of range contraction -- will presage range expansion. You can think of a stocks movement like a spring / slinky -- it need to recoil (contract) in order to build up enough potential energy for its next expansion. However, it's important thing to note that range contraction doesn't tell you the direction of the impending expansion only that an expansion may be coming. It's up to the trader to use other means (indicators, common sense?) to catch the expansion in the right direction.

One of the nice things about trading against narrow range bars is that they give you a tight stop. If you size your positions like I do, the smaller the stop means you can buy (or short) more shares. It's the trades with the really tight stops that let you make those huge R-Multiple trades that often make or break a trader. Here's an example from a trade I took last week.

I'm going to use the actual entries taken by myself and by Butterboy but instead of using my actual exits, I'll assume I held until the close. So let's say my equity is $100,000 and my R, the amount I want to risk per trade, is 1% or $1,000. Given that, here's the first trade, based off of my actual entry parameters:

• Entry at 53.96
• Stop was 50 cents higher so I could short 1000/.50 or 2,000 shares
• DGX closed at 49.64, which is $4.32 lower. So that trade returned $8,640 or 8.64 times my risk (8.64R)

Here's the DGX trade using Butterboy's entry criteria. He used a 15 minute chart and entered lower than I did but his stop was also much tighter:

• Entry at 53.40
• I believe his stop was 21 cents higher, just above the 11:30 bar which he entered against. Given that one could short 1,000/.21 or just over 4,700 shares.
• DGX closed at $3.76 beneath the entry so that trade returned $17,672 or 17.67 times the initial risk (17.64R).

It should be clear why so many of us like narrow range bars and why I scan for them. That second trade had a worse (lower) entry but because of its tighter stop it was almost twice as profitable. It also shows why I love the percent risk position sizing model.

Here's an article on range contraction written by Oswald S. Castillo, TheStockStalker and very possibly the #1 fan of range contraction trading. Oswald also has a webinar on his site entitled "Principles of Trading Range Contraction". It's a good video but you have to register to view it.

You may also find Alan Farley's work on "coiled springs" and NR7s helpful. He goes in-depth on coiled springs in his book, The Master Swing Trader.

P.S. I almost forgot to answer the question of: How Narrow is Narrow?

For daytrading I like to find candles that are 1% or less than the price of the stock. So for a $20 stock any candles with a range of 20 cents or less really get my attention. In cases like those I just need small moves in the stock to rack up some nice gains. I'll typically pass on candles with a range much greater than 1% because they'll require a much larger move in the stock in order for me to get those large R-multiples.

For swing trading (looking at daily charts) I don't have a concrete answer on how narrow is narrow. But you could probably come up with some value based on the stock's average true range (ATR) or some other method. For example, you could define narrow as less than half a stocks ATR.

NRx (narrowest range on the last X candles) is always a good way to go as well, and it's easy to scan for.

Michelle B submits: I remember reading about a French housewife who when asked by her husband every morning what they would be having for dinner would always give the same tireless reply: "It depends on what jumps into my market basket." Some traders prefer to trade from a prepared watchlist. For me, trading is interesting because I do not know what will jump into my market basket. My preparation is my confidence in my ability to execute perceived opportunities according to my risk parameters.

Using the premarket top gainers/losers scan, I search the NASDAQ, NYSE, and AMEX market stalls for any tasty morsels and look for any stocks moving on the highs/lows scan. I also check Briefing.com and MarketWatch Newsfinder for news that has happened since the close of the previous trading day---identifying which stocks the market considers to be newsworthy is one of the many trading skills which can be developed.

In addition, I note via Briefing.com calendars such events as economic reports, earnings, Fedspeak, splits, upgrades/downgrades, and conferences. Usually, I have around ten to twelve trading candidates. Sometimes, there are none or a very small number. In that case, I patiently wait until the market has a new batch of fresh candidates---I will not make do with stale merchandise.

I list the stock symbols of my candidates and the important economic events for that day in my journal. Next to each stock symbol, I put the results of my basic Yahoo financial research---reason for price movement, float size, and short interest. I identify support and resistance on various timeframes. Sometimes, I check Yahoo Finance message boards and blogs to find out why a stock may be moving. Though in that case, I am always on my guard to weed the chaff from the wheat.

Then, I place my candidates on the chart pages of my Advanced Trading Platform (ATP). Four thirty-minute charts go on pages two through four. The main page contains the following windows: order entry, order status, real time portfolio, high/low scan, top gainers/losers scan, watchlist containing, in addition to my trading candidates, the major indices and technical indicators like VIX and TICK. Last but not least, is a chart window for the most promising candidate.

Discovering that ININ---set to gap up because it has given higher earnings' guidance---has a very small float and moderately high short interest, coupled with my knowledge that such stocks often move very strongly on low volume days, like Fridays, pre-holidays, and summer days, makes ININ the star attraction at this stage.

The market opens, and since I am particularly interested in seeing if ININ can trigger a trade, it gets top attention with a one-minute chart (above) on my main trading page. Meanwhile, I am flipping through my chart pages to see if anything is close to triggering---nothing yet. I check the one-minute chart of ININ, and I see the choppy pennant being formed at 10 A.M. The pennant range is around .25. That's enough of a sign for me to buy half of my lot as close to the bottom of this formation at 13.65. Buying a partial lot before the actual breakout happens makes sense in high-probability trades, especially if the desired lot size is substantial. I am set to buy the rest via an automatic buy stop if it can clear the top of the pennant at 13.87 on the 5 minute (see below). Trades moving strongly in the morning require monitoring via shorter time frames.

Fifteen minutes later, it does, and my full lot is in, average cost around 13.78. Just before the breakout, the pennant was around seventy-five percent completed and the volume dried up. Triangles/pennants need to be not completely formed or else they usually just fizzle out and do not support strong breakouts. The pause in volume is what my trading buddy calls a lullipop, a lull before the breakout, where the equilibrium between buyers and sellers end and the buyers take over.

Part of my prepurchase research of ININ was its support and resistance based on weekly price level resistance. Since the open of the July 24, 2006 weekly, bearish, high-volume candle was 14.81 (see below), therefore trapping lots of traders at that level, I determined the resistance to be around 14.80. The target of 14.80 also matches the length of the pole on which the flag formed on Friday's intraday charts, forming a measured move and confirming the resistance.

As the ININ trade continues, I am monitoring the price action on different timeframes, and have my mental stop under the flag for the first half lot at 13.53 and then just above the flag at 13.89 for the complete lot. Price is smoothly rising---each candle opens near but not with much overlapping to the close of the previous candle. I put my sell limit in for the full lot just under the target price. My offer gets lifted fairly easily, and I am out with a $1.00 move gotten in about 40 minutes. Since the risk was .25 for the complete lot, it was a +4R trade. For quickly moving stocks, I will often forgo hard stops, but it is a risk because my net connection could be interrupted. Then my carefully constructed trade will exist only in cyberspace floating without my pilotage. So, I am working on the artful science of consistently setting hard regular/trailing stops with upper triggers via bracketed orders to take care of that possibility.

My market basket is now empty until the next fresh thing jumps into it.

There seems to be a lot of controversial over the concept of R-Multiples. I've been seeing people complain about them for months now and I've been meaning to write a post about "R". I really wanted to do it last week but I'm glad I didn't get to it because this week a raging debate about R has popped up. Glenn, at DehTrader can serve as the poster child for the anti-R crew. Here's part of his recent rant against R-Multiples (emphasis is mine):

I post real numbers as opposed to R values, I always have. I like real numbers, I understand real numbers and I see truth in real numbers and I think the reader does too. As a reader of many blogs I find zero value in any post or summaries containing R values, I don't see any point in sharing that information. I suppose if I posted in R values I could look like a pretty good trader, but we all know I am a struggling trader. R can mean anything so why even bother with it... R stands for bullsh!t imo and that's my rant (that and ads haha). The best blogs out there post real numbers, Boogtser, JC (NYSE), the Kirkster all come to mind.

He's joined by folks like Paul who left this comment over on Ugly's post about R multiples:

I believe dollar values are more important than R value. I agree that the actual $ value is meaningless. However, R values are subjective and don’t give you a true idea on how successful the trade was. If you defined your risk at 15 cents and made 30 cents on the trade, while another person made 50 cents but decided his risk would be 50 cents, R values would say the guy who made 30 cents was more successful. I have a problem with that. It could very well be that the guy who only risked 15 cents is playing it too safe and his 2R gain was a bad trade.

So that gives you an idea of the anti-R sentiment. I'm going to explain why I think R-Multiples are so useful and why I use them in my trading and on this site.

What is R?

R is simply the dollar risk per trade. It's nothing but a reward-to-risk ratio. I first heard it called "R" in Van Tharp's book "Trade Your Way to Financial Freedom". In another of his books, "Financial Freedom Through Electronic Day Trading", Dr. Tharp reveals the great secret of trading:

The golden rule of trading is to keep losses at a level of 1 R as often as possible and to make profits that are high-R multiples.

You often hear (read) that traders should only look for trades with a reward/risk ratio of at least 2 or 3 to 1. Expressing your results in terms of how many times your risk allows you to easily see how well your trades measure up to such a standard. So when I look at my results in terms of multiples of R I can easily tell how good or bad the trades were. I like to think of R-Multiples as telling you the efficiency of your system.

So why not just use dollars?

Expressing my results in dollars would achieve the same result if I always risked the same amount of money. But what if I triple my account and therefore trade larger positions compared to when I started trading? Or what if I hit a rough spot and decide to cut my share size down while I ride out the storm? Then the dollar results won't easily tell me how trades from one period of time compared to another period of time. But if I use R making such comparisons is simple. Either my trades passed the risk / reward ratio test or they didn't. The actual number of dollars at risk doesn't matter, how many multiples of the dollars at risk does.

Along the same lines, recording trades in terms of R-Multiples allows you to easily calculate your system's expectancy. (Follow the link for why you should care about expectancy.)

Also, as Rx said:

talking and thinking in terms of R-multiple when you discuss about profits is an excellent approach - that by itself makes you focus on risk and money management - the actual "grail" to successful trading.

That is a very important point. Whenever I see people posting dollar returns, especially losses, that are all over the place the first thing I ask myself is "I wonder what his risk per trade is". It's almost a certainty that those traders aren't focusing on risk and as a result keep having huge losses. The mere fact that you have to define R and then place a stop to keep your loss to 1R is probably too constraining for those gamblers traders. Dr. Tharp says about determining your initial stop-loss point as soon as you enter a trade, which, by definition woud give you a 1 R loss:

This principle is so important that if you cannot follow it, then you might as well give up the idea of electronic day trading right away.

The reason I use R on the blog is because I don't want to discuss dollars or my account size on the site (as Ugly stated). That's nobody's business but mine. Also, it makes it easy for people to figure out what they could have made or lost on a trade with their own account size and risk per trade amount. If you see a trade that returned 3R all one has to do is plug in their dollar risk per trade to figure out what they could have made / lost.

To the R Haters

Let me address the "alleged" issues which I quoted above...

Glenn thinks that R is just some made up number and could mean anything. He likes "real" numbers. While it may be fun to see that somebody made $10,000 on a trade that in and of itself doesn't tell you how good that trade was. What if that person risked $30,000 to make that $10,000? Or what if they risked $1,000 to make that $10,000? Those are two very different trades. Sure they both made the same amount of money but isn't the second trade a much more efficient use of capital?

What if somebody is trading $500,000 lots to make $1,000 in profit? It may be nice to see somebody saying that they made $1,000 here and $1,000 there but damn(!) that's an inefficient use of capital. So while R could mean anything in terms of dollars, in my humble opinion what really matters is how many multiples of R were made or lost. That tells you the quality of a trade or system.

Glenn also states that if he reported his trades in terms of R he could appear to be a good trader. I'm sorry to tell him that's simply not the case. If you lost money that means your expectancy, which is just your average return expressed in R-Multiples, was negative.

Paul said that "R values are subjective and don’t give you a true idea on how successful the trade was". That is exactly wrong. R-multiples are the very thing that tells you exactly how successful a given trade was, if you choose to grade on a risk/reward basis.

Percentages vs. Dollars

This debate about R reminds me of a conversation I had a couple of weeks ago. I was in a presentation for Trade-Ideas' new tool, the Odds Maker. They were showing how you could backtest all these different scenarios with the tool. The results were expressed in average dollars won or lost. Another viewer and I asked about seeing the results in percentages. They kept saying that perhaps they would do that in a later revision. I kept harping on it because to me seeing the results in dollars was of little use for the way I size my positions

The argument from the presenter was that all you had to do was multiply the average dollar return by your average lot size to figure out how much money you could have made with a given system you were testing. I had to disagree because my lot size can vary drastically depending on how far away my stop loss is. Here's a situation which could be problematic -- I trade Google with a 2 point stop (which is only about half of a percent) and get lucky and make 6 points of profit. All of my other trades are on stocks under $50 with stops less than 50 cents. I could have some combination of winners and losers mixed in there... most of them probably well less than $6. That $6 gain may skew the results when presented as average dollars won. That's an over-simplification and there are all kinds of possible permutations. But I hope my point is clear that looking at the results in terms of average dollars won/lost may not tell accurately tell you the story.

So how can we make the results clearer? Simple, express them in percentages. That way, regardless of how many shares were traded or the prices of the stocks traded the results can be equalized across all the trades. I feel much better being able to say , "OK, this system would have returned X%" instead of "X number of points.

We debated the merits of each way of reporting for a few minutes and at one point somebody said, well , for this release we're aiming for the "lowest common denominator". In other words, the average person can't think in percentages, so we're just gonna report in points. I was like, F the average person, make it work the "right" way! The funny thing is that after debating all of that the software actually could express the results in percentage terms. We just had to switch a setting.

So my point of that little story is that I always prefer to think in terms of percentages in stead of points. I always see people talking about number of shares of point moves. For example, you might hear somebody exclaim "Google is up 5 points!!!" I don't see that as anything to get excited about. That just over a 1% move -- a normal fluctuation. You'll hear similar things from reporters talking excitedly about the Dow being up some triple-digit amount. The Nasdaq may actually be up a lot more on a percentage basis but they'll just say, eh, the Nasdaq is "just" up 30 points.

Looking at the percentages makes those kind of comparisons easier. R-Multiples do the same thing for traders. They can accurately compare their own trades and they can take another trader's results expressed in R and easily relate them to their own system.

Cameron just left a question for me in the comments to last night's recap. He asked:

I'm looking at your way of thinking about the market. One of the biggest questions I have, though, is regarding your stops.

You said you set your stop at just .20 over your short dollar amount. With all of the intra-day fluctuations, how can you know that it wouldn't have gone up, say, .25 and then nose-dived? It's a perfectly plausible situation.

Said another way, I can place a buy order with a stop, but we easily have 1% swings in a day. If a swing occurs, I have automatically sold for a lower price, only to see the market go back up. Therefore, I'm encouraged to set much broader stops. You set yours for only .20 over the buy price, though, which is why I'm confused.

I often get asked variations of this question, so I thought I'd put my answer in its own post. This was my response:

I couldn't have known that the stock wouldn't have shaken me out. But I also wouldn't have known if I'd have been shaken out if my stop was 50 cents or even a dollar higher. That narrow range candle just provided me with a tight stop. I like tight stops b/c they give me a better opportunity to get high R-multiple winners.

But to your point, how tight of a stop is too tight and will practically guarantee being stopped out? It's hard to answer that question but I generally shy away from any stop that's less than 0.5% of the stock price. I normally like my stops to be about 1% of the stock price. In general, I'm more comfortable with somewhat tighter stops on the QQQQ b/c of how liquid it is and b/c it's less volatile than an "in play" stock would be.

Also, note that I didn't pick a 20 cent stop out of thin air. I went with what the previous candle told me was resistance.

The key thing to remember is that I'm on the hunt for high R-Multiple winners. Because of the way I size my positions (the percent risk model), the tighter my stops the higher the potential reward. In other words if I think a stock has potential to move $2 I'd rather have a 20 cent stop than a $1 stop. Risking 20 cents to make 200 cents ($2) would give me a 10R winner vs only a 2R winner if I had risked $1. (If R was 1% of my equity that would be a 10% gain vs a 2% gain.) For a real example of this see the DGX trades I posted on my Narrow Range Bars post.

So there' a trade-off at work here between stop size and risk/reward multiples. The tighter the stop, the better chance you have of getting a high R winner but you also have a better chance of getting stopped out. Your win ratio is likely to be inversely proportional to how tight your stops are -- looser stops should lead to a higher win rate but also fewer high R-multiple winners.

It's also key that you don't pick your stops willy-nilly. I always use the support/resistance levels that the charts dictate to me.

There's some good advice in Alan Farley's latest article to help "home gamers" (Cramer's phrase) survive the current extra-volatile market. The whole article is worth reading but here are the four bits of advice:

I have four pieces of cautionary advice for my at-home brethren. Listen up, because it's no longer a question of whether you want to take real money out of the market, or just add a few bucks to the weekly shopping budget. These remedial steps must be taken if you want to survive long enough to take advantage of the real opportunity.

1. Wait for the Market Volatility Index (VIX) to drop below 40 and stay there for a week. Massive price swings require equally massive stop losses, which rarely justify the intended positions. Your only alternative is to stand aside and do nothing, no matter how much it hurts to watch others playing those big rallies and selloffs.
2. Forget overnight positions until the index futures stop gapping 2% or more every morning. These price jolts are great news when you're on the right side of the trade, but total devastation if you're on the wrong side. And guess what? You're not smart enough to predict overnight direction from day to day. Neither am I.
3. End your love affair with popular stocks that made you money during the last bull market. In November 2008, these are the issues that will trigger the most painful and unexpected reversals, which happen right after you're absolutely convinced your position is the right play. The bottom line: They see you coming, sucker.
4. Get control of the time element in your market strategy. You're getting killed because you have no patience and forgot how to sit on your hands when your trading edge isn't in play. Realistically, it could be months before the market works for you again. Would you rather wait it out and survive, or stay busy and get crushed?

My thoughts:

1. I like that part about watching the VIX. A few weeks ago I decided to cut my risk per trade (R) in half (from 1% to 0.5% of my equity) until the VIX breaks that top it's trying to put in. That means a close below 44. (Dr. Brett recently wrote about when traders should cut their risk.)
2. Tough to argue with what he said about overnight holds. This is certainly a market for intraday trades.
3. Reminds me of what I said about AAPL and GOOG back in July. Add DRYS and its ilk to the list too. It's better to just have an open mind and grab on to whichever stocks become the new leaders.
4. Yep, preservation of capital must be the top priority. If you can't sit out at least lower your risk.

I was recently asked about how I determine my "universe" of stocks. That is the stocks that I consider fair game to be traded. I use different filters for swing trading and day trading but they do have similar criteria. In both cases I'm trying to find liquid, volatile stocks. Once the filter's applied I then look for certain patterns / setups to trade.

For swing trade candidates I use the following criteria in TeleChart:

• Price Per Share: about $5 to Max
• 5 Day Average Volume: 75th percentile & higher (right now that works out to about 466,000 shares per day)
• Price Volatility: 25th percentile & higher
• Volume (Dollars) 5-Day: 75th percentile & higher

Here's a shot of the filter. The count column shows how many stocks are left after applying each row in the filter.

That typically narrows the market down to about 1,000 stocks. Then I look for setups among that set of stocks. I gave a lot of details about the setups I scan for back in the day.

For day trading I'm trying to find liquid stocks which are abnormally active and volatile. To do that I use the following two scans in Trade-Ideas:

Bearish Scanner's Filter:

• Min Price: $10
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: -0.5% (negative means down at least 0.5%)

Bullish Scanner's Filter:

• Min Price: $5
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: 1%

Sometimes I'll tweak these settings to give more or fewer alerts. I usually do that by adjusting the "Min Current Volume". Details about the actual alerts & setups I track for daytrading are on my "how I work" page. If you use Trade-Ideas you can paste the following link into the "Collaborate" feature and load that filter and all the alerts I track:

Bearish -- grab the URL of this link

Bullish -- grab the URL of this link

One thing I have always preached on the blog, less so on Twitter, where I bang out more trading ideas and market thoughts is that shorting stocks is hard. I think it’s harder than any aspect of learning the market. It’s dangerous. Let this morning be the only reminder you should EVER need.

Howard's right about shorting being difficult. I think it's so hard because it's not simply the opposite of going long. (I won't even go into the whole thing about your losses while short being theoretically infinite. Been there... done that. Use a stop to limit your losses, use proper position sizing, ONLY short LIQUID stocks and you'll be fine.) What makes shorting tricky is that bear moves often have violent (short-covering) rallies because the psychology of the crowd trading a down market is different than that of a bull market. You have to be quick on your feet when shorting. My motto is "stick & move".

Many traders love to buy breakouts in bull markets. (Whether that's actually a good strategy is debatable). In my experience swing trading, the opposite of that strategy, shorting breakdowns through resistance, will often lead you right into a snapback rally and, as MaoXian used to say "the quickest loss ever". That's why I often make note of all the people who are initiating shorts after the market has already fallen to a major support level. We saw that this week as I noted in some of the morning watchlists.

My contention is that if you were caught short Friday morning you should consider your losses as tuition paid to the school of hard knocks. Learn from that expensive lesson, take your losses and hopefully survive to trade another day. I'll stop short of saying that the losses were deserved but there were plenty of warnings to at least cover your shorts if not to get long. There are so many good sources of market information these days, both on the web and, yes, even on TV. I'm not saying to blindly follow somebody else's opinion but it can be helpful to see what others are doing based on what they see. Here are just a few of the recent warning signs:

• T2108 dropping below 20 -- Ah, good old reliable T2108. I've been watching it closely as we've sold off. On Wednesday I noted that it finally hit the point where wise shorts would want to cover. It pays to find a good overbought/oversold indicator and heed its warnings. (You may need different indicators or settings for different timeframes.) Sell at overbought and cover at oversold. A couple of years ago I decided to force myself to put some IRA money to work every time T2108 broke 20. It hasn't failed me yet.
• The Fed (Plunge Protection Team) has interfered announced stimulus packages around the July lows a couple times. On Thursday and earlier in the week there was talk of more PPT action.
• The PPT took action earlier in the week and last week. Just look at the orchestration of the LEH and AIG situations, etc.
• Extreme Volatility -- The moves all week were nothing short of violent. Positions were whipsawed all over the place. That in & of itself would be reason enough to lighten up on positions if not move to the sidelines. That kind of volatility is often a sign of a trend reversal, not of a continuation of the previous trend. That's why so many traders watch the VIX. Tons of people noted the spike in the VIX on Thursday.
• Corey from the 'Afraid to Trade' blog warned 'Use Extreme Caution in the Week Ahead.' He gave many good reasons, including the Federal Reserve interest rate decision, the quadruple witching options expiration and headline risk from troubled financial firms. His crystal ball was working well when he wrote "We could see a week ahead that will be discussed years later - as such, if you are a newer trader, it might be best to switch to simulation mode this week or use this week as a training experience, rather than risking real capital in an environment that could swing violently up and down due to market events scheduled to happen this week."
• Bullish Technical Divergences -- Dr. Brett pointed out some bullish divergences he was seeing in the market. Perhaps most important were what he called 'those fuzzy indicators' -- "Traffic on the blog is way up, reflecting trader uncertainty and desire for information. I just fielded my fourth media interview request in two days. During quiet and bullish market periods, I don't get four requests in a month." On Tuesday morning I also noted my blog's traffic spiked as did Barry Ritholtz. I've often joked about making some kind of sentiment indicator based on my site's traffic ebb & flow & referral logs. It's not a bad idea and I think it would be especially useful if it were based on a major financial site's traffic data.
• Dennis Gartman telling folks to "be small" -- Gartman was on CNBC's 'Fast Money' early in the week saying that he was scared of the market's movement and he was "being small" and planned to "get smaller". His advice to others was to "be small" in this market.
• Jeff Macke, also on 'Fast Money' was warning people not to play (trade) if they didn't understand the (changing) rules of the game. He was referring to all the headline risk from PPT action and the volatility caused by rumors -- many of which were spread by CNBC during the trading sessions. Ironically, Macke was kicking himself Thursday night for not following his own advice and getting caught short.

I fully believe that had the PPT not acted on Thursday night the market was *destined* to move higher in the short term on its own. Still being short on expiration Friday in this environment was just asking for trouble. So what's a trader to do? Like I said earlier, stick & move. I think it makes much more sense to short bounces back to a trendline or moving average. William O'Neil's book on short selling talks about using retracements to the 50 and/or 200-day moving average as a place to initiate shorts. By the time the stock (or whatever instrument you're trading) is extended from its moving average it's time to cover. Then you can wait for another bounce to reload.

The strategy to cover & reload makes more sense when you're short due to the fact that if you're dead right with your call on the stock dropping the most you can gain is 100%. The odds of that happening are slim and if it ever did happen you'd likely have to ride out some severe short squeezes. But you might be able to stick & move your way to more that a 100% gain by reloading multiple times. In theory, you could catch 15 10% moves lower in a stock that's falling, retracing & falling anew.

Short sellers need to be nimble, pick their entry and exit spots wisely and heed signs of impending reversals. Trying to short breakdowns of an already extended market is a sucker's play -- as is overstaying your welcome while short.

I got this question the other day:

When you short a stock at 30.00 and you want to place a stop limit in case it jumps up past your max loss of say, 31.00, how exactly is that order entered (everytime I try to do this, it sells (Ed: he means buys) my stock immediately - I'm doing something really wrong). Do I put a limit price of 31.00 and a stop price above that, i.e. 31.50 (thus saying it's ok to repurchase any shares between 31.00 and 31.50)?

This is one of the trickier orders to place. In that situation you would want to enter a buy stop limit order (buy on stop with a limit). The stop would be at 31.00 and the limit would be 31.50. Here's a screen shot of a similar order to cover QQQQ with a buy stop @ 47 with a limit of 47.10:

What would happen here is that once the stop price of 47.00 is hit a buy limit order of 47.10 would be entered. As with any limit order, you're guaranteed price but not execution, so you may not be filled. If you wanted to be sure that you got out (covered) just use a buy stop of $47.00.

In case you missed it in the links section the other day, be sure to read the interview with Michelle B over on the StockTickr blog.

The other day Downtown Trader asked me how I handled trailing my stop on that DGX short. I pretty much followed exactly what's in my post about how I take partial profits. I moved the stop to breakeven once my gain was equal to my initial risk of 50 cents per share. As it dropped further I moved the stop to lock in one-third of my profit. Why one-third? It's a value I came up with which I think allows a stock to have a normal retracement without stopping me out. It's sort of an arbitrary level, although I got the idea from the Fibonacci levels. I've found that method to work reasonably well after much trial and error with trailing stops. I read somewhere that you can consider a retracement to be normal as long as it didn't break the 62% Fib level. So I figured that I could trail my stops somewhere around that level to allow stocks to have their normal fluctuations. I generally get more aggressive with the trailing stops as the day wears on. I'll move them to 50% by 3:00 and just keep pushing them until the close -- assuming I don't bail before then like I did with DGX.

The other day I got the following email about narrow range bars:

I was going through your site and others and I keep on hearing, no entry on narrow range bars. What is a narrow range bar, and could you send me a picture of one?

A narrow range bar (candle) is simply a bar which has a range from high to low that's much less than the average bar for a given equity. You'll often see people, like Dr. Sen, tracking NR7 bars. NR7 means the narrowest range of the last seven bars. For example, here are some NR7s from last Friday:

So now that you know what narrow range candles are, why should you care about them? Since stocks generally cycle between periods of low volatility and high volatility it can be beneficial to identify when the high volatility times are approaching. Often times, especially for trending charts, narrow range periods -- periods of range contraction -- will presage range expansion. You can think of a stocks movement like a spring / slinky -- it need to recoil (contract) in order to build up enough potential energy for its next expansion. However, it's important thing to note that range contraction doesn't tell you the direction of the impending expansion only that an expansion may be coming. It's up to the trader to use other means (indicators, common sense?) to catch the expansion in the right direction.

One of the nice things about trading against narrow range bars is that they give you a tight stop. If you size your positions like I do, the smaller the stop means you can buy (or short) more shares. It's the trades with the really tight stops that let you make those huge R-Multiple trades that often make or break a trader. Here's an example from a trade I took last week.

I'm going to use the actual entries taken by myself and by Butterboy but instead of using my actual exits, I'll assume I held until the close. So let's say my equity is $100,000 and my R, the amount I want to risk per trade, is 1% or $1,000. Given that, here's the first trade, based off of my actual entry parameters:

• Entry at 53.96
• Stop was 50 cents higher so I could short 1000/.50 or 2,000 shares
• DGX closed at 49.64, which is $4.32 lower. So that trade returned $8,640 or 8.64 times my risk (8.64R)

Here's the DGX trade using Butterboy's entry criteria. He used a 15 minute chart and entered lower than I did but his stop was also much tighter:

• Entry at 53.40
• I believe his stop was 21 cents higher, just above the 11:30 bar which he entered against. Given that one could short 1,000/.21 or just over 4,700 shares.
• DGX closed at $3.76 beneath the entry so that trade returned $17,672 or 17.67 times the initial risk (17.64R).

It should be clear why so many of us like narrow range bars and why I scan for them. That second trade had a worse (lower) entry but because of its tighter stop it was almost twice as profitable. It also shows why I love the percent risk position sizing model.

Here's an article on range contraction written by Oswald S. Castillo, TheStockStalker and very possibly the #1 fan of range contraction trading. Oswald also has a webinar on his site entitled "Principles of Trading Range Contraction". It's a good video but you have to register to view it.

You may also find Alan Farley's work on "coiled springs" and NR7s helpful. He goes in-depth on coiled springs in his book, The Master Swing Trader.

P.S. I almost forgot to answer the question of: How Narrow is Narrow?

For daytrading I like to find candles that are 1% or less than the price of the stock. So for a $20 stock any candles with a range of 20 cents or less really get my attention. In cases like those I just need small moves in the stock to rack up some nice gains. I'll typically pass on candles with a range much greater than 1% because they'll require a much larger move in the stock in order for me to get those large R-multiples.

For swing trading (looking at daily charts) I don't have a concrete answer on how narrow is narrow. But you could probably come up with some value based on the stock's average true range (ATR) or some other method. For example, you could define narrow as less than half a stocks ATR.

NRx (narrowest range on the last X candles) is always a good way to go as well, and it's easy to scan for.

Michelle B submits: I remember reading about a French housewife who when asked by her husband every morning what they would be having for dinner would always give the same tireless reply: "It depends on what jumps into my market basket." Some traders prefer to trade from a prepared watchlist. For me, trading is interesting because I do not know what will jump into my market basket. My preparation is my confidence in my ability to execute perceived opportunities according to my risk parameters.

Using the premarket top gainers/losers scan, I search the NASDAQ, NYSE, and AMEX market stalls for any tasty morsels and look for any stocks moving on the highs/lows scan. I also check Briefing.com and MarketWatch Newsfinder for news that has happened since the close of the previous trading day---identifying which stocks the market considers to be newsworthy is one of the many trading skills which can be developed.

In addition, I note via Briefing.com calendars such events as economic reports, earnings, Fedspeak, splits, upgrades/downgrades, and conferences. Usually, I have around ten to twelve trading candidates. Sometimes, there are none or a very small number. In that case, I patiently wait until the market has a new batch of fresh candidates---I will not make do with stale merchandise.

I list the stock symbols of my candidates and the important economic events for that day in my journal. Next to each stock symbol, I put the results of my basic Yahoo financial research---reason for price movement, float size, and short interest. I identify support and resistance on various timeframes. Sometimes, I check Yahoo Finance message boards and blogs to find out why a stock may be moving. Though in that case, I am always on my guard to weed the chaff from the wheat.

Then, I place my candidates on the chart pages of my Advanced Trading Platform (ATP). Four thirty-minute charts go on pages two through four. The main page contains the following windows: order entry, order status, real time portfolio, high/low scan, top gainers/losers scan, watchlist containing, in addition to my trading candidates, the major indices and technical indicators like VIX and TICK. Last but not least, is a chart window for the most promising candidate.

Discovering that ININ---set to gap up because it has given higher earnings' guidance---has a very small float and moderately high short interest, coupled with my knowledge that such stocks often move very strongly on low volume days, like Fridays, pre-holidays, and summer days, makes ININ the star attraction at this stage.

The market opens, and since I am particularly interested in seeing if ININ can trigger a trade, it gets top attention with a one-minute chart (above) on my main trading page. Meanwhile, I am flipping through my chart pages to see if anything is close to triggering---nothing yet. I check the one-minute chart of ININ, and I see the choppy pennant being formed at 10 A.M. The pennant range is around .25. That's enough of a sign for me to buy half of my lot as close to the bottom of this formation at 13.65. Buying a partial lot before the actual breakout happens makes sense in high-probability trades, especially if the desired lot size is substantial. I am set to buy the rest via an automatic buy stop if it can clear the top of the pennant at 13.87 on the 5 minute (see below). Trades moving strongly in the morning require monitoring via shorter time frames.

Fifteen minutes later, it does, and my full lot is in, average cost around 13.78. Just before the breakout, the pennant was around seventy-five percent completed and the volume dried up. Triangles/pennants need to be not completely formed or else they usually just fizzle out and do not support strong breakouts. The pause in volume is what my trading buddy calls a lullipop, a lull before the breakout, where the equilibrium between buyers and sellers end and the buyers take over.

Part of my prepurchase research of ININ was its support and resistance based on weekly price level resistance. Since the open of the July 24, 2006 weekly, bearish, high-volume candle was 14.81 (see below), therefore trapping lots of traders at that level, I determined the resistance to be around 14.80. The target of 14.80 also matches the length of the pole on which the flag formed on Friday's intraday charts, forming a measured move and confirming the resistance.

As the ININ trade continues, I am monitoring the price action on different timeframes, and have my mental stop under the flag for the first half lot at 13.53 and then just above the flag at 13.89 for the complete lot. Price is smoothly rising---each candle opens near but not with much overlapping to the close of the previous candle. I put my sell limit in for the full lot just under the target price. My offer gets lifted fairly easily, and I am out with a $1.00 move gotten in about 40 minutes. Since the risk was .25 for the complete lot, it was a +4R trade. For quickly moving stocks, I will often forgo hard stops, but it is a risk because my net connection could be interrupted. Then my carefully constructed trade will exist only in cyberspace floating without my pilotage. So, I am working on the artful science of consistently setting hard regular/trailing stops with upper triggers via bracketed orders to take care of that possibility.

My market basket is now empty until the next fresh thing jumps into it.

There seems to be a lot of controversial over the concept of R-Multiples. I've been seeing people complain about them for months now and I've been meaning to write a post about "R". I really wanted to do it last week but I'm glad I didn't get to it because this week a raging debate about R has popped up. Glenn, at DehTrader can serve as the poster child for the anti-R crew. Here's part of his recent rant against R-Multiples (emphasis is mine):

I post real numbers as opposed to R values, I always have. I like real numbers, I understand real numbers and I see truth in real numbers and I think the reader does too. As a reader of many blogs I find zero value in any post or summaries containing R values, I don't see any point in sharing that information. I suppose if I posted in R values I could look like a pretty good trader, but we all know I am a struggling trader. R can mean anything so why even bother with it... R stands for bullsh!t imo and that's my rant (that and ads haha). The best blogs out there post real numbers, Boogtser, JC (NYSE), the Kirkster all come to mind.

He's joined by folks like Paul who left this comment over on Ugly's post about R multiples:

I believe dollar values are more important than R value. I agree that the actual $ value is meaningless. However, R values are subjective and don’t give you a true idea on how successful the trade was. If you defined your risk at 15 cents and made 30 cents on the trade, while another person made 50 cents but decided his risk would be 50 cents, R values would say the guy who made 30 cents was more successful. I have a problem with that. It could very well be that the guy who only risked 15 cents is playing it too safe and his 2R gain was a bad trade.

So that gives you an idea of the anti-R sentiment. I'm going to explain why I think R-Multiples are so useful and why I use them in my trading and on this site.

What is R?

R is simply the dollar risk per trade. It's nothing but a reward-to-risk ratio. I first heard it called "R" in Van Tharp's book "Trade Your Way to Financial Freedom". In another of his books, "Financial Freedom Through Electronic Day Trading", Dr. Tharp reveals the great secret of trading:

The golden rule of trading is to keep losses at a level of 1 R as often as possible and to make profits that are high-R multiples.

You often hear (read) that traders should only look for trades with a reward/risk ratio of at least 2 or 3 to 1. Expressing your results in terms of how many times your risk allows you to easily see how well your trades measure up to such a standard. So when I look at my results in terms of multiples of R I can easily tell how good or bad the trades were. I like to think of R-Multiples as telling you the efficiency of your system.

So why not just use dollars?

Expressing my results in dollars would achieve the same result if I always risked the same amount of money. But what if I triple my account and therefore trade larger positions compared to when I started trading? Or what if I hit a rough spot and decide to cut my share size down while I ride out the storm? Then the dollar results won't easily tell me how trades from one period of time compared to another period of time. But if I use R making such comparisons is simple. Either my trades passed the risk / reward ratio test or they didn't. The actual number of dollars at risk doesn't matter, how many multiples of the dollars at risk does.

Along the same lines, recording trades in terms of R-Multiples allows you to easily calculate your system's expectancy. (Follow the link for why you should care about expectancy.)

Also, as Rx said:

talking and thinking in terms of R-multiple when you discuss about profits is an excellent approach - that by itself makes you focus on risk and money management - the actual "grail" to successful trading.

That is a very important point. Whenever I see people posting dollar returns, especially losses, that are all over the place the first thing I ask myself is "I wonder what his risk per trade is". It's almost a certainty that those traders aren't focusing on risk and as a result keep having huge losses. The mere fact that you have to define R and then place a stop to keep your loss to 1R is probably too constraining for those gamblers traders. Dr. Tharp says about determining your initial stop-loss point as soon as you enter a trade, which, by definition woud give you a 1 R loss:

This principle is so important that if you cannot follow it, then you might as well give up the idea of electronic day trading right away.

The reason I use R on the blog is because I don't want to discuss dollars or my account size on the site (as Ugly stated). That's nobody's business but mine. Also, it makes it easy for people to figure out what they could have made or lost on a trade with their own account size and risk per trade amount. If you see a trade that returned 3R all one has to do is plug in their dollar risk per trade to figure out what they could have made / lost.

To the R Haters

Let me address the "alleged" issues which I quoted above...

Glenn thinks that R is just some made up number and could mean anything. He likes "real" numbers. While it may be fun to see that somebody made $10,000 on a trade that in and of itself doesn't tell you how good that trade was. What if that person risked $30,000 to make that $10,000? Or what if they risked $1,000 to make that $10,000? Those are two very different trades. Sure they both made the same amount of money but isn't the second trade a much more efficient use of capital?

What if somebody is trading $500,000 lots to make $1,000 in profit? It may be nice to see somebody saying that they made $1,000 here and $1,000 there but damn(!) that's an inefficient use of capital. So while R could mean anything in terms of dollars, in my humble opinion what really matters is how many multiples of R were made or lost. That tells you the quality of a trade or system.

Glenn also states that if he reported his trades in terms of R he could appear to be a good trader. I'm sorry to tell him that's simply not the case. If you lost money that means your expectancy, which is just your average return expressed in R-Multiples, was negative.

Paul said that "R values are subjective and don’t give you a true idea on how successful the trade was". That is exactly wrong. R-multiples are the very thing that tells you exactly how successful a given trade was, if you choose to grade on a risk/reward basis.

Percentages vs. Dollars

This debate about R reminds me of a conversation I had a couple of weeks ago. I was in a presentation for Trade-Ideas' new tool, the Odds Maker. They were showing how you could backtest all these different scenarios with the tool. The results were expressed in average dollars won or lost. Another viewer and I asked about seeing the results in percentages. They kept saying that perhaps they would do that in a later revision. I kept harping on it because to me seeing the results in dollars was of little use for the way I size my positions

The argument from the presenter was that all you had to do was multiply the average dollar return by your average lot size to figure out how much money you could have made with a given system you were testing. I had to disagree because my lot size can vary drastically depending on how far away my stop loss is. Here's a situation which could be problematic -- I trade Google with a 2 point stop (which is only about half of a percent) and get lucky and make 6 points of profit. All of my other trades are on stocks under $50 with stops less than 50 cents. I could have some combination of winners and losers mixed in there... most of them probably well less than $6. That $6 gain may skew the results when presented as average dollars won. That's an over-simplification and there are all kinds of possible permutations. But I hope my point is clear that looking at the results in terms of average dollars won/lost may not tell accurately tell you the story.

So how can we make the results clearer? Simple, express them in percentages. That way, regardless of how many shares were traded or the prices of the stocks traded the results can be equalized across all the trades. I feel much better being able to say , "OK, this system would have returned X%" instead of "X number of points.

We debated the merits of each way of reporting for a few minutes and at one point somebody said, well , for this release we're aiming for the "lowest common denominator". In other words, the average person can't think in percentages, so we're just gonna report in points. I was like, F the average person, make it work the "right" way! The funny thing is that after debating all of that the software actually could express the results in percentage terms. We just had to switch a setting.

So my point of that little story is that I always prefer to think in terms of percentages in stead of points. I always see people talking about number of shares of point moves. For example, you might hear somebody exclaim "Google is up 5 points!!!" I don't see that as anything to get excited about. That just over a 1% move -- a normal fluctuation. You'll hear similar things from reporters talking excitedly about the Dow being up some triple-digit amount. The Nasdaq may actually be up a lot more on a percentage basis but they'll just say, eh, the Nasdaq is "just" up 30 points.

Looking at the percentages makes those kind of comparisons easier. R-Multiples do the same thing for traders. They can accurately compare their own trades and they can take another trader's results expressed in R and easily relate them to their own system.

Cameron just left a question for me in the comments to last night's recap. He asked:

I'm looking at your way of thinking about the market. One of the biggest questions I have, though, is regarding your stops.

You said you set your stop at just .20 over your short dollar amount. With all of the intra-day fluctuations, how can you know that it wouldn't have gone up, say, .25 and then nose-dived? It's a perfectly plausible situation.

Said another way, I can place a buy order with a stop, but we easily have 1% swings in a day. If a swing occurs, I have automatically sold for a lower price, only to see the market go back up. Therefore, I'm encouraged to set much broader stops. You set yours for only .20 over the buy price, though, which is why I'm confused.

I often get asked variations of this question, so I thought I'd put my answer in its own post. This was my response:

I couldn't have known that the stock wouldn't have shaken me out. But I also wouldn't have known if I'd have been shaken out if my stop was 50 cents or even a dollar higher. That narrow range candle just provided me with a tight stop. I like tight stops b/c they give me a better opportunity to get high R-multiple winners.

But to your point, how tight of a stop is too tight and will practically guarantee being stopped out? It's hard to answer that question but I generally shy away from any stop that's less than 0.5% of the stock price. I normally like my stops to be about 1% of the stock price. In general, I'm more comfortable with somewhat tighter stops on the QQQQ b/c of how liquid it is and b/c it's less volatile than an "in play" stock would be.

Also, note that I didn't pick a 20 cent stop out of thin air. I went with what the previous candle told me was resistance.

The key thing to remember is that I'm on the hunt for high R-Multiple winners. Because of the way I size my positions (the percent risk model), the tighter my stops the higher the potential reward. In other words if I think a stock has potential to move $2 I'd rather have a 20 cent stop than a $1 stop. Risking 20 cents to make 200 cents ($2) would give me a 10R winner vs only a 2R winner if I had risked $1. (If R was 1% of my equity that would be a 10% gain vs a 2% gain.) For a real example of this see the DGX trades I posted on my Narrow Range Bars post.

So there' a trade-off at work here between stop size and risk/reward multiples. The tighter the stop, the better chance you have of getting a high R winner but you also have a better chance of getting stopped out. Your win ratio is likely to be inversely proportional to how tight your stops are -- looser stops should lead to a higher win rate but also fewer high R-multiple winners.

It's also key that you don't pick your stops willy-nilly. I always use the support/resistance levels that the charts dictate to me.

There's some good advice in Alan Farley's latest article to help "home gamers" (Cramer's phrase) survive the current extra-volatile market. The whole article is worth reading but here are the four bits of advice:

I have four pieces of cautionary advice for my at-home brethren. Listen up, because it's no longer a question of whether you want to take real money out of the market, or just add a few bucks to the weekly shopping budget. These remedial steps must be taken if you want to survive long enough to take advantage of the real opportunity.

1. Wait for the Market Volatility Index (VIX) to drop below 40 and stay there for a week. Massive price swings require equally massive stop losses, which rarely justify the intended positions. Your only alternative is to stand aside and do nothing, no matter how much it hurts to watch others playing those big rallies and selloffs.
2. Forget overnight positions until the index futures stop gapping 2% or more every morning. These price jolts are great news when you're on the right side of the trade, but total devastation if you're on the wrong side. And guess what? You're not smart enough to predict overnight direction from day to day. Neither am I.
3. End your love affair with popular stocks that made you money during the last bull market. In November 2008, these are the issues that will trigger the most painful and unexpected reversals, which happen right after you're absolutely convinced your position is the right play. The bottom line: They see you coming, sucker.
4. Get control of the time element in your market strategy. You're getting killed because you have no patience and forgot how to sit on your hands when your trading edge isn't in play. Realistically, it could be months before the market works for you again. Would you rather wait it out and survive, or stay busy and get crushed?

My thoughts:

1. I like that part about watching the VIX. A few weeks ago I decided to cut my risk per trade (R) in half (from 1% to 0.5% of my equity) until the VIX breaks that top it's trying to put in. That means a close below 44. (Dr. Brett recently wrote about when traders should cut their risk.)
2. Tough to argue with what he said about overnight holds. This is certainly a market for intraday trades.
3. Reminds me of what I said about AAPL and GOOG back in July. Add DRYS and its ilk to the list too. It's better to just have an open mind and grab on to whichever stocks become the new leaders.
4. Yep, preservation of capital must be the top priority. If you can't sit out at least lower your risk.

I was recently asked about how I determine my "universe" of stocks. That is the stocks that I consider fair game to be traded. I use different filters for swing trading and day trading but they do have similar criteria. In both cases I'm trying to find liquid, volatile stocks. Once the filter's applied I then look for certain patterns / setups to trade.

For swing trade candidates I use the following criteria in TeleChart:

• Price Per Share: about $5 to Max
• 5 Day Average Volume: 75th percentile & higher (right now that works out to about 466,000 shares per day)
• Price Volatility: 25th percentile & higher
• Volume (Dollars) 5-Day: 75th percentile & higher

Here's a shot of the filter. The count column shows how many stocks are left after applying each row in the filter.

That typically narrows the market down to about 1,000 stocks. Then I look for setups among that set of stocks. I gave a lot of details about the setups I scan for back in the day.

For day trading I'm trying to find liquid stocks which are abnormally active and volatile. To do that I use the following two scans in Trade-Ideas:

Bearish Scanner's Filter:

• Min Price: $10
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: -0.5% (negative means down at least 0.5%)

Bullish Scanner's Filter:

• Min Price: $5
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: 1%

Sometimes I'll tweak these settings to give more or fewer alerts. I usually do that by adjusting the "Min Current Volume". Details about the actual alerts & setups I track for daytrading are on my "how I work" page. If you use Trade-Ideas you can paste the following link into the "Collaborate" feature and load that filter and all the alerts I track:

Bearish -- grab the URL of this link

Bullish -- grab the URL of this link

One thing I have always preached on the blog, less so on Twitter, where I bang out more trading ideas and market thoughts is that shorting stocks is hard. I think it’s harder than any aspect of learning the market. It’s dangerous. Let this morning be the only reminder you should EVER need.

Howard's right about shorting being difficult. I think it's so hard because it's not simply the opposite of going long. (I won't even go into the whole thing about your losses while short being theoretically infinite. Been there... done that. Use a stop to limit your losses, use proper position sizing, ONLY short LIQUID stocks and you'll be fine.) What makes shorting tricky is that bear moves often have violent (short-covering) rallies because the psychology of the crowd trading a down market is different than that of a bull market. You have to be quick on your feet when shorting. My motto is "stick & move".

Many traders love to buy breakouts in bull markets. (Whether that's actually a good strategy is debatable). In my experience swing trading, the opposite of that strategy, shorting breakdowns through resistance, will often lead you right into a snapback rally and, as MaoXian used to say "the quickest loss ever". That's why I often make note of all the people who are initiating shorts after the market has already fallen to a major support level. We saw that this week as I noted in some of the morning watchlists.

My contention is that if you were caught short Friday morning you should consider your losses as tuition paid to the school of hard knocks. Learn from that expensive lesson, take your losses and hopefully survive to trade another day. I'll stop short of saying that the losses were deserved but there were plenty of warnings to at least cover your shorts if not to get long. There are so many good sources of market information these days, both on the web and, yes, even on TV. I'm not saying to blindly follow somebody else's opinion but it can be helpful to see what others are doing based on what they see. Here are just a few of the recent warning signs:

• T2108 dropping below 20 -- Ah, good old reliable T2108. I've been watching it closely as we've sold off. On Wednesday I noted that it finally hit the point where wise shorts would want to cover. It pays to find a good overbought/oversold indicator and heed its warnings. (You may need different indicators or settings for different timeframes.) Sell at overbought and cover at oversold. A couple of years ago I decided to force myself to put some IRA money to work every time T2108 broke 20. It hasn't failed me yet.
• The Fed (Plunge Protection Team) has interfered announced stimulus packages around the July lows a couple times. On Thursday and earlier in the week there was talk of more PPT action.
• The PPT took action earlier in the week and last week. Just look at the orchestration of the LEH and AIG situations, etc.
• Extreme Volatility -- The moves all week were nothing short of violent. Positions were whipsawed all over the place. That in & of itself would be reason enough to lighten up on positions if not move to the sidelines. That kind of volatility is often a sign of a trend reversal, not of a continuation of the previous trend. That's why so many traders watch the VIX. Tons of people noted the spike in the VIX on Thursday.
• Corey from the 'Afraid to Trade' blog warned 'Use Extreme Caution in the Week Ahead.' He gave many good reasons, including the Federal Reserve interest rate decision, the quadruple witching options expiration and headline risk from troubled financial firms. His crystal ball was working well when he wrote "We could see a week ahead that will be discussed years later - as such, if you are a newer trader, it might be best to switch to simulation mode this week or use this week as a training experience, rather than risking real capital in an environment that could swing violently up and down due to market events scheduled to happen this week."
• Bullish Technical Divergences -- Dr. Brett pointed out some bullish divergences he was seeing in the market. Perhaps most important were what he called 'those fuzzy indicators' -- "Traffic on the blog is way up, reflecting trader uncertainty and desire for information. I just fielded my fourth media interview request in two days. During quiet and bullish market periods, I don't get four requests in a month." On Tuesday morning I also noted my blog's traffic spiked as did Barry Ritholtz. I've often joked about making some kind of sentiment indicator based on my site's traffic ebb & flow & referral logs. It's not a bad idea and I think it would be especially useful if it were based on a major financial site's traffic data.
• Dennis Gartman telling folks to "be small" -- Gartman was on CNBC's 'Fast Money' early in the week saying that he was scared of the market's movement and he was "being small" and planned to "get smaller". His advice to others was to "be small" in this market.
• Jeff Macke, also on 'Fast Money' was warning people not to play (trade) if they didn't understand the (changing) rules of the game. He was referring to all the headline risk from PPT action and the volatility caused by rumors -- many of which were spread by CNBC during the trading sessions. Ironically, Macke was kicking himself Thursday night for not following his own advice and getting caught short.

I fully believe that had the PPT not acted on Thursday night the market was *destined* to move higher in the short term on its own. Still being short on expiration Friday in this environment was just asking for trouble. So what's a trader to do? Like I said earlier, stick & move. I think it makes much more sense to short bounces back to a trendline or moving average. William O'Neil's book on short selling talks about using retracements to the 50 and/or 200-day moving average as a place to initiate shorts. By the time the stock (or whatever instrument you're trading) is extended from its moving average it's time to cover. Then you can wait for another bounce to reload.

The strategy to cover & reload makes more sense when you're short due to the fact that if you're dead right with your call on the stock dropping the most you can gain is 100%. The odds of that happening are slim and if it ever did happen you'd likely have to ride out some severe short squeezes. But you might be able to stick & move your way to more that a 100% gain by reloading multiple times. In theory, you could catch 15 10% moves lower in a stock that's falling, retracing & falling anew.

Short sellers need to be nimble, pick their entry and exit spots wisely and heed signs of impending reversals. Trying to short breakdowns of an already extended market is a sucker's play -- as is overstaying your welcome while short.

I got this question the other day:

When you short a stock at 30.00 and you want to place a stop limit in case it jumps up past your max loss of say, 31.00, how exactly is that order entered (everytime I try to do this, it sells (Ed: he means buys) my stock immediately - I'm doing something really wrong). Do I put a limit price of 31.00 and a stop price above that, i.e. 31.50 (thus saying it's ok to repurchase any shares between 31.00 and 31.50)?

This is one of the trickier orders to place. In that situation you would want to enter a buy stop limit order (buy on stop with a limit). The stop would be at 31.00 and the limit would be 31.50. Here's a screen shot of a similar order to cover QQQQ with a buy stop @ 47 with a limit of 47.10:

What would happen here is that once the stop price of 47.00 is hit a buy limit order of 47.10 would be entered. As with any limit order, you're guaranteed price but not execution, so you may not be filled. If you wanted to be sure that you got out (covered) just use a buy stop of $47.00.

In case you missed it in the links section the other day, be sure to read the interview with Michelle B over on the StockTickr blog.

The other day Downtown Trader asked me how I handled trailing my stop on that DGX short. I pretty much followed exactly what's in my post about how I take partial profits. I moved the stop to breakeven once my gain was equal to my initial risk of 50 cents per share. As it dropped further I moved the stop to lock in one-third of my profit. Why one-third? It's a value I came up with which I think allows a stock to have a normal retracement without stopping me out. It's sort of an arbitrary level, although I got the idea from the Fibonacci levels. I've found that method to work reasonably well after much trial and error with trailing stops. I read somewhere that you can consider a retracement to be normal as long as it didn't break the 62% Fib level. So I figured that I could trail my stops somewhere around that level to allow stocks to have their normal fluctuations. I generally get more aggressive with the trailing stops as the day wears on. I'll move them to 50% by 3:00 and just keep pushing them until the close -- assuming I don't bail before then like I did with DGX.

The other day I got the following email about narrow range bars:

I was going through your site and others and I keep on hearing, no entry on narrow range bars. What is a narrow range bar, and could you send me a picture of one?

A narrow range bar (candle) is simply a bar which has a range from high to low that's much less than the average bar for a given equity. You'll often see people, like Dr. Sen, tracking NR7 bars. NR7 means the narrowest range of the last seven bars. For example, here are some NR7s from last Friday:

So now that you know what narrow range candles are, why should you care about them? Since stocks generally cycle between periods of low volatility and high volatility it can be beneficial to identify when the high volatility times are approaching. Often times, especially for trending charts, narrow range periods -- periods of range contraction -- will presage range expansion. You can think of a stocks movement like a spring / slinky -- it need to recoil (contract) in order to build up enough potential energy for its next expansion. However, it's important thing to note that range contraction doesn't tell you the direction of the impending expansion only that an expansion may be coming. It's up to the trader to use other means (indicators, common sense?) to catch the expansion in the right direction.

One of the nice things about trading against narrow range bars is that they give you a tight stop. If you size your positions like I do, the smaller the stop means you can buy (or short) more shares. It's the trades with the really tight stops that let you make those huge R-Multiple trades that often make or break a trader. Here's an example from a trade I took last week.

I'm going to use the actual entries taken by myself and by Butterboy but instead of using my actual exits, I'll assume I held until the close. So let's say my equity is $100,000 and my R, the amount I want to risk per trade, is 1% or $1,000. Given that, here's the first trade, based off of my actual entry parameters:

• Entry at 53.96
• Stop was 50 cents higher so I could short 1000/.50 or 2,000 shares
• DGX closed at 49.64, which is $4.32 lower. So that trade returned $8,640 or 8.64 times my risk (8.64R)

Here's the DGX trade using Butterboy's entry criteria. He used a 15 minute chart and entered lower than I did but his stop was also much tighter:

• Entry at 53.40
• I believe his stop was 21 cents higher, just above the 11:30 bar which he entered against. Given that one could short 1,000/.21 or just over 4,700 shares.
• DGX closed at $3.76 beneath the entry so that trade returned $17,672 or 17.67 times the initial risk (17.64R).

It should be clear why so many of us like narrow range bars and why I scan for them. That second trade had a worse (lower) entry but because of its tighter stop it was almost twice as profitable. It also shows why I love the percent risk position sizing model.

Here's an article on range contraction written by Oswald S. Castillo, TheStockStalker and very possibly the #1 fan of range contraction trading. Oswald also has a webinar on his site entitled "Principles of Trading Range Contraction". It's a good video but you have to register to view it.

You may also find Alan Farley's work on "coiled springs" and NR7s helpful. He goes in-depth on coiled springs in his book, The Master Swing Trader.

P.S. I almost forgot to answer the question of: How Narrow is Narrow?

For daytrading I like to find candles that are 1% or less than the price of the stock. So for a $20 stock any candles with a range of 20 cents or less really get my attention. In cases like those I just need small moves in the stock to rack up some nice gains. I'll typically pass on candles with a range much greater than 1% because they'll require a much larger move in the stock in order for me to get those large R-multiples.

For swing trading (looking at daily charts) I don't have a concrete answer on how narrow is narrow. But you could probably come up with some value based on the stock's average true range (ATR) or some other method. For example, you could define narrow as less than half a stocks ATR.

NRx (narrowest range on the last X candles) is always a good way to go as well, and it's easy to scan for.

Michelle B submits: I remember reading about a French housewife who when asked by her husband every morning what they would be having for dinner would always give the same tireless reply: "It depends on what jumps into my market basket." Some traders prefer to trade from a prepared watchlist. For me, trading is interesting because I do not know what will jump into my market basket. My preparation is my confidence in my ability to execute perceived opportunities according to my risk parameters.

Using the premarket top gainers/losers scan, I search the NASDAQ, NYSE, and AMEX market stalls for any tasty morsels and look for any stocks moving on the highs/lows scan. I also check Briefing.com and MarketWatch Newsfinder for news that has happened since the close of the previous trading day---identifying which stocks the market considers to be newsworthy is one of the many trading skills which can be developed.

In addition, I note via Briefing.com calendars such events as economic reports, earnings, Fedspeak, splits, upgrades/downgrades, and conferences. Usually, I have around ten to twelve trading candidates. Sometimes, there are none or a very small number. In that case, I patiently wait until the market has a new batch of fresh candidates---I will not make do with stale merchandise.

I list the stock symbols of my candidates and the important economic events for that day in my journal. Next to each stock symbol, I put the results of my basic Yahoo financial research---reason for price movement, float size, and short interest. I identify support and resistance on various timeframes. Sometimes, I check Yahoo Finance message boards and blogs to find out why a stock may be moving. Though in that case, I am always on my guard to weed the chaff from the wheat.

Then, I place my candidates on the chart pages of my Advanced Trading Platform (ATP). Four thirty-minute charts go on pages two through four. The main page contains the following windows: order entry, order status, real time portfolio, high/low scan, top gainers/losers scan, watchlist containing, in addition to my trading candidates, the major indices and technical indicators like VIX and TICK. Last but not least, is a chart window for the most promising candidate.

Discovering that ININ---set to gap up because it has given higher earnings' guidance---has a very small float and moderately high short interest, coupled with my knowledge that such stocks often move very strongly on low volume days, like Fridays, pre-holidays, and summer days, makes ININ the star attraction at this stage.

The market opens, and since I am particularly interested in seeing if ININ can trigger a trade, it gets top attention with a one-minute chart (above) on my main trading page. Meanwhile, I am flipping through my chart pages to see if anything is close to triggering---nothing yet. I check the one-minute chart of ININ, and I see the choppy pennant being formed at 10 A.M. The pennant range is around .25. That's enough of a sign for me to buy half of my lot as close to the bottom of this formation at 13.65. Buying a partial lot before the actual breakout happens makes sense in high-probability trades, especially if the desired lot size is substantial. I am set to buy the rest via an automatic buy stop if it can clear the top of the pennant at 13.87 on the 5 minute (see below). Trades moving strongly in the morning require monitoring via shorter time frames.

Fifteen minutes later, it does, and my full lot is in, average cost around 13.78. Just before the breakout, the pennant was around seventy-five percent completed and the volume dried up. Triangles/pennants need to be not completely formed or else they usually just fizzle out and do not support strong breakouts. The pause in volume is what my trading buddy calls a lullipop, a lull before the breakout, where the equilibrium between buyers and sellers end and the buyers take over.

Part of my prepurchase research of ININ was its support and resistance based on weekly price level resistance. Since the open of the July 24, 2006 weekly, bearish, high-volume candle was 14.81 (see below), therefore trapping lots of traders at that level, I determined the resistance to be around 14.80. The target of 14.80 also matches the length of the pole on which the flag formed on Friday's intraday charts, forming a measured move and confirming the resistance.

As the ININ trade continues, I am monitoring the price action on different timeframes, and have my mental stop under the flag for the first half lot at 13.53 and then just above the flag at 13.89 for the complete lot. Price is smoothly rising---each candle opens near but not with much overlapping to the close of the previous candle. I put my sell limit in for the full lot just under the target price. My offer gets lifted fairly easily, and I am out with a $1.00 move gotten in about 40 minutes. Since the risk was .25 for the complete lot, it was a +4R trade. For quickly moving stocks, I will often forgo hard stops, but it is a risk because my net connection could be interrupted. Then my carefully constructed trade will exist only in cyberspace floating without my pilotage. So, I am working on the artful science of consistently setting hard regular/trailing stops with upper triggers via bracketed orders to take care of that possibility.

My market basket is now empty until the next fresh thing jumps into it.

There seems to be a lot of controversial over the concept of R-Multiples. I've been seeing people complain about them for months now and I've been meaning to write a post about "R". I really wanted to do it last week but I'm glad I didn't get to it because this week a raging debate about R has popped up. Glenn, at DehTrader can serve as the poster child for the anti-R crew. Here's part of his recent rant against R-Multiples (emphasis is mine):

I post real numbers as opposed to R values, I always have. I like real numbers, I understand real numbers and I see truth in real numbers and I think the reader does too. As a reader of many blogs I find zero value in any post or summaries containing R values, I don't see any point in sharing that information. I suppose if I posted in R values I could look like a pretty good trader, but we all know I am a struggling trader. R can mean anything so why even bother with it... R stands for bullsh!t imo and that's my rant (that and ads haha). The best blogs out there post real numbers, Boogtser, JC (NYSE), the Kirkster all come to mind.

He's joined by folks like Paul who left this comment over on Ugly's post about R multiples:

I believe dollar values are more important than R value. I agree that the actual $ value is meaningless. However, R values are subjective and don’t give you a true idea on how successful the trade was. If you defined your risk at 15 cents and made 30 cents on the trade, while another person made 50 cents but decided his risk would be 50 cents, R values would say the guy who made 30 cents was more successful. I have a problem with that. It could very well be that the guy who only risked 15 cents is playing it too safe and his 2R gain was a bad trade.

So that gives you an idea of the anti-R sentiment. I'm going to explain why I think R-Multiples are so useful and why I use them in my trading and on this site.

What is R?

R is simply the dollar risk per trade. It's nothing but a reward-to-risk ratio. I first heard it called "R" in Van Tharp's book "Trade Your Way to Financial Freedom". In another of his books, "Financial Freedom Through Electronic Day Trading", Dr. Tharp reveals the great secret of trading:

The golden rule of trading is to keep losses at a level of 1 R as often as possible and to make profits that are high-R multiples.

You often hear (read) that traders should only look for trades with a reward/risk ratio of at least 2 or 3 to 1. Expressing your results in terms of how many times your risk allows you to easily see how well your trades measure up to such a standard. So when I look at my results in terms of multiples of R I can easily tell how good or bad the trades were. I like to think of R-Multiples as telling you the efficiency of your system.

So why not just use dollars?

Expressing my results in dollars would achieve the same result if I always risked the same amount of money. But what if I triple my account and therefore trade larger positions compared to when I started trading? Or what if I hit a rough spot and decide to cut my share size down while I ride out the storm? Then the dollar results won't easily tell me how trades from one period of time compared to another period of time. But if I use R making such comparisons is simple. Either my trades passed the risk / reward ratio test or they didn't. The actual number of dollars at risk doesn't matter, how many multiples of the dollars at risk does.

Along the same lines, recording trades in terms of R-Multiples allows you to easily calculate your system's expectancy. (Follow the link for why you should care about expectancy.)

Also, as Rx said:

talking and thinking in terms of R-multiple when you discuss about profits is an excellent approach - that by itself makes you focus on risk and money management - the actual "grail" to successful trading.

That is a very important point. Whenever I see people posting dollar returns, especially losses, that are all over the place the first thing I ask myself is "I wonder what his risk per trade is". It's almost a certainty that those traders aren't focusing on risk and as a result keep having huge losses. The mere fact that you have to define R and then place a stop to keep your loss to 1R is probably too constraining for those gamblers traders. Dr. Tharp says about determining your initial stop-loss point as soon as you enter a trade, which, by definition woud give you a 1 R loss:

This principle is so important that if you cannot follow it, then you might as well give up the idea of electronic day trading right away.

The reason I use R on the blog is because I don't want to discuss dollars or my account size on the site (as Ugly stated). That's nobody's business but mine. Also, it makes it easy for people to figure out what they could have made or lost on a trade with their own account size and risk per trade amount. If you see a trade that returned 3R all one has to do is plug in their dollar risk per trade to figure out what they could have made / lost.

To the R Haters

Let me address the "alleged" issues which I quoted above...

Glenn thinks that R is just some made up number and could mean anything. He likes "real" numbers. While it may be fun to see that somebody made $10,000 on a trade that in and of itself doesn't tell you how good that trade was. What if that person risked $30,000 to make that $10,000? Or what if they risked $1,000 to make that $10,000? Those are two very different trades. Sure they both made the same amount of money but isn't the second trade a much more efficient use of capital?

What if somebody is trading $500,000 lots to make $1,000 in profit? It may be nice to see somebody saying that they made $1,000 here and $1,000 there but damn(!) that's an inefficient use of capital. So while R could mean anything in terms of dollars, in my humble opinion what really matters is how many multiples of R were made or lost. That tells you the quality of a trade or system.

Glenn also states that if he reported his trades in terms of R he could appear to be a good trader. I'm sorry to tell him that's simply not the case. If you lost money that means your expectancy, which is just your average return expressed in R-Multiples, was negative.

Paul said that "R values are subjective and don’t give you a true idea on how successful the trade was". That is exactly wrong. R-multiples are the very thing that tells you exactly how successful a given trade was, if you choose to grade on a risk/reward basis.

Percentages vs. Dollars

This debate about R reminds me of a conversation I had a couple of weeks ago. I was in a presentation for Trade-Ideas' new tool, the Odds Maker. They were showing how you could backtest all these different scenarios with the tool. The results were expressed in average dollars won or lost. Another viewer and I asked about seeing the results in percentages. They kept saying that perhaps they would do that in a later revision. I kept harping on it because to me seeing the results in dollars was of little use for the way I size my positions

The argument from the presenter was that all you had to do was multiply the average dollar return by your average lot size to figure out how much money you could have made with a given system you were testing. I had to disagree because my lot size can vary drastically depending on how far away my stop loss is. Here's a situation which could be problematic -- I trade Google with a 2 point stop (which is only about half of a percent) and get lucky and make 6 points of profit. All of my other trades are on stocks under $50 with stops less than 50 cents. I could have some combination of winners and losers mixed in there... most of them probably well less than $6. That $6 gain may skew the results when presented as average dollars won. That's an over-simplification and there are all kinds of possible permutations. But I hope my point is clear that looking at the results in terms of average dollars won/lost may not tell accurately tell you the story.

So how can we make the results clearer? Simple, express them in percentages. That way, regardless of how many shares were traded or the prices of the stocks traded the results can be equalized across all the trades. I feel much better being able to say , "OK, this system would have returned X%" instead of "X number of points.

We debated the merits of each way of reporting for a few minutes and at one point somebody said, well , for this release we're aiming for the "lowest common denominator". In other words, the average person can't think in percentages, so we're just gonna report in points. I was like, F the average person, make it work the "right" way! The funny thing is that after debating all of that the software actually could express the results in percentage terms. We just had to switch a setting.

So my point of that little story is that I always prefer to think in terms of percentages in stead of points. I always see people talking about number of shares of point moves. For example, you might hear somebody exclaim "Google is up 5 points!!!" I don't see that as anything to get excited about. That just over a 1% move -- a normal fluctuation. You'll hear similar things from reporters talking excitedly about the Dow being up some triple-digit amount. The Nasdaq may actually be up a lot more on a percentage basis but they'll just say, eh, the Nasdaq is "just" up 30 points.

Looking at the percentages makes those kind of comparisons easier. R-Multiples do the same thing for traders. They can accurately compare their own trades and they can take another trader's results expressed in R and easily relate them to their own system.

Cameron just left a question for me in the comments to last night's recap. He asked:

I'm looking at your way of thinking about the market. One of the biggest questions I have, though, is regarding your stops.

You said you set your stop at just .20 over your short dollar amount. With all of the intra-day fluctuations, how can you know that it wouldn't have gone up, say, .25 and then nose-dived? It's a perfectly plausible situation.

Said another way, I can place a buy order with a stop, but we easily have 1% swings in a day. If a swing occurs, I have automatically sold for a lower price, only to see the market go back up. Therefore, I'm encouraged to set much broader stops. You set yours for only .20 over the buy price, though, which is why I'm confused.

I often get asked variations of this question, so I thought I'd put my answer in its own post. This was my response:

I couldn't have known that the stock wouldn't have shaken me out. But I also wouldn't have known if I'd have been shaken out if my stop was 50 cents or even a dollar higher. That narrow range candle just provided me with a tight stop. I like tight stops b/c they give me a better opportunity to get high R-multiple winners.

But to your point, how tight of a stop is too tight and will practically guarantee being stopped out? It's hard to answer that question but I generally shy away from any stop that's less than 0.5% of the stock price. I normally like my stops to be about 1% of the stock price. In general, I'm more comfortable with somewhat tighter stops on the QQQQ b/c of how liquid it is and b/c it's less volatile than an "in play" stock would be.

Also, note that I didn't pick a 20 cent stop out of thin air. I went with what the previous candle told me was resistance.

The key thing to remember is that I'm on the hunt for high R-Multiple winners. Because of the way I size my positions (the percent risk model), the tighter my stops the higher the potential reward. In other words if I think a stock has potential to move $2 I'd rather have a 20 cent stop than a $1 stop. Risking 20 cents to make 200 cents ($2) would give me a 10R winner vs only a 2R winner if I had risked $1. (If R was 1% of my equity that would be a 10% gain vs a 2% gain.) For a real example of this see the DGX trades I posted on my Narrow Range Bars post.

So there' a trade-off at work here between stop size and risk/reward multiples. The tighter the stop, the better chance you have of getting a high R winner but you also have a better chance of getting stopped out. Your win ratio is likely to be inversely proportional to how tight your stops are -- looser stops should lead to a higher win rate but also fewer high R-multiple winners.

It's also key that you don't pick your stops willy-nilly. I always use the support/resistance levels that the charts dictate to me.

There's some good advice in Alan Farley's latest article to help "home gamers" (Cramer's phrase) survive the current extra-volatile market. The whole article is worth reading but here are the four bits of advice:

I have four pieces of cautionary advice for my at-home brethren. Listen up, because it's no longer a question of whether you want to take real money out of the market, or just add a few bucks to the weekly shopping budget. These remedial steps must be taken if you want to survive long enough to take advantage of the real opportunity.

1. Wait for the Market Volatility Index (VIX) to drop below 40 and stay there for a week. Massive price swings require equally massive stop losses, which rarely justify the intended positions. Your only alternative is to stand aside and do nothing, no matter how much it hurts to watch others playing those big rallies and selloffs.
2. Forget overnight positions until the index futures stop gapping 2% or more every morning. These price jolts are great news when you're on the right side of the trade, but total devastation if you're on the wrong side. And guess what? You're not smart enough to predict overnight direction from day to day. Neither am I.
3. End your love affair with popular stocks that made you money during the last bull market. In November 2008, these are the issues that will trigger the most painful and unexpected reversals, which happen right after you're absolutely convinced your position is the right play. The bottom line: They see you coming, sucker.
4. Get control of the time element in your market strategy. You're getting killed because you have no patience and forgot how to sit on your hands when your trading edge isn't in play. Realistically, it could be months before the market works for you again. Would you rather wait it out and survive, or stay busy and get crushed?

My thoughts:

1. I like that part about watching the VIX. A few weeks ago I decided to cut my risk per trade (R) in half (from 1% to 0.5% of my equity) until the VIX breaks that top it's trying to put in. That means a close below 44. (Dr. Brett recently wrote about when traders should cut their risk.)
2. Tough to argue with what he said about overnight holds. This is certainly a market for intraday trades.
3. Reminds me of what I said about AAPL and GOOG back in July. Add DRYS and its ilk to the list too. It's better to just have an open mind and grab on to whichever stocks become the new leaders.
4. Yep, preservation of capital must be the top priority. If you can't sit out at least lower your risk.

I was recently asked about how I determine my "universe" of stocks. That is the stocks that I consider fair game to be traded. I use different filters for swing trading and day trading but they do have similar criteria. In both cases I'm trying to find liquid, volatile stocks. Once the filter's applied I then look for certain patterns / setups to trade.

For swing trade candidates I use the following criteria in TeleChart:

• Price Per Share: about $5 to Max
• 5 Day Average Volume: 75th percentile & higher (right now that works out to about 466,000 shares per day)
• Price Volatility: 25th percentile & higher
• Volume (Dollars) 5-Day: 75th percentile & higher

Here's a shot of the filter. The count column shows how many stocks are left after applying each row in the filter.

That typically narrows the market down to about 1,000 stocks. Then I look for setups among that set of stocks. I gave a lot of details about the setups I scan for back in the day.

For day trading I'm trying to find liquid stocks which are abnormally active and volatile. To do that I use the following two scans in Trade-Ideas:

Bearish Scanner's Filter:

• Min Price: $10
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: -0.5% (negative means down at least 0.5%)

Bullish Scanner's Filter:

• Min Price: $5
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: 1%

Sometimes I'll tweak these settings to give more or fewer alerts. I usually do that by adjusting the "Min Current Volume". Details about the actual alerts & setups I track for daytrading are on my "how I work" page. If you use Trade-Ideas you can paste the following link into the "Collaborate" feature and load that filter and all the alerts I track:

Bearish -- grab the URL of this link

Bullish -- grab the URL of this link

One thing I have always preached on the blog, less so on Twitter, where I bang out more trading ideas and market thoughts is that shorting stocks is hard. I think it’s harder than any aspect of learning the market. It’s dangerous. Let this morning be the only reminder you should EVER need.

Howard's right about shorting being difficult. I think it's so hard because it's not simply the opposite of going long. (I won't even go into the whole thing about your losses while short being theoretically infinite. Been there... done that. Use a stop to limit your losses, use proper position sizing, ONLY short LIQUID stocks and you'll be fine.) What makes shorting tricky is that bear moves often have violent (short-covering) rallies because the psychology of the crowd trading a down market is different than that of a bull market. You have to be quick on your feet when shorting. My motto is "stick & move".

Many traders love to buy breakouts in bull markets. (Whether that's actually a good strategy is debatable). In my experience swing trading, the opposite of that strategy, shorting breakdowns through resistance, will often lead you right into a snapback rally and, as MaoXian used to say "the quickest loss ever". That's why I often make note of all the people who are initiating shorts after the market has already fallen to a major support level. We saw that this week as I noted in some of the morning watchlists.

My contention is that if you were caught short Friday morning you should consider your losses as tuition paid to the school of hard knocks. Learn from that expensive lesson, take your losses and hopefully survive to trade another day. I'll stop short of saying that the losses were deserved but there were plenty of warnings to at least cover your shorts if not to get long. There are so many good sources of market information these days, both on the web and, yes, even on TV. I'm not saying to blindly follow somebody else's opinion but it can be helpful to see what others are doing based on what they see. Here are just a few of the recent warning signs:

• T2108 dropping below 20 -- Ah, good old reliable T2108. I've been watching it closely as we've sold off. On Wednesday I noted that it finally hit the point where wise shorts would want to cover. It pays to find a good overbought/oversold indicator and heed its warnings. (You may need different indicators or settings for different timeframes.) Sell at overbought and cover at oversold. A couple of years ago I decided to force myself to put some IRA money to work every time T2108 broke 20. It hasn't failed me yet.
• The Fed (Plunge Protection Team) has interfered announced stimulus packages around the July lows a couple times. On Thursday and earlier in the week there was talk of more PPT action.
• The PPT took action earlier in the week and last week. Just look at the orchestration of the LEH and AIG situations, etc.
• Extreme Volatility -- The moves all week were nothing short of violent. Positions were whipsawed all over the place. That in & of itself would be reason enough to lighten up on positions if not move to the sidelines. That kind of volatility is often a sign of a trend reversal, not of a continuation of the previous trend. That's why so many traders watch the VIX. Tons of people noted the spike in the VIX on Thursday.
• Corey from the 'Afraid to Trade' blog warned 'Use Extreme Caution in the Week Ahead.' He gave many good reasons, including the Federal Reserve interest rate decision, the quadruple witching options expiration and headline risk from troubled financial firms. His crystal ball was working well when he wrote "We could see a week ahead that will be discussed years later - as such, if you are a newer trader, it might be best to switch to simulation mode this week or use this week as a training experience, rather than risking real capital in an environment that could swing violently up and down due to market events scheduled to happen this week."
• Bullish Technical Divergences -- Dr. Brett pointed out some bullish divergences he was seeing in the market. Perhaps most important were what he called 'those fuzzy indicators' -- "Traffic on the blog is way up, reflecting trader uncertainty and desire for information. I just fielded my fourth media interview request in two days. During quiet and bullish market periods, I don't get four requests in a month." On Tuesday morning I also noted my blog's traffic spiked as did Barry Ritholtz. I've often joked about making some kind of sentiment indicator based on my site's traffic ebb & flow & referral logs. It's not a bad idea and I think it would be especially useful if it were based on a major financial site's traffic data.
• Dennis Gartman telling folks to "be small" -- Gartman was on CNBC's 'Fast Money' early in the week saying that he was scared of the market's movement and he was "being small" and planned to "get smaller". His advice to others was to "be small" in this market.
• Jeff Macke, also on 'Fast Money' was warning people not to play (trade) if they didn't understand the (changing) rules of the game. He was referring to all the headline risk from PPT action and the volatility caused by rumors -- many of which were spread by CNBC during the trading sessions. Ironically, Macke was kicking himself Thursday night for not following his own advice and getting caught short.

I fully believe that had the PPT not acted on Thursday night the market was *destined* to move higher in the short term on its own. Still being short on expiration Friday in this environment was just asking for trouble. So what's a trader to do? Like I said earlier, stick & move. I think it makes much more sense to short bounces back to a trendline or moving average. William O'Neil's book on short selling talks about using retracements to the 50 and/or 200-day moving average as a place to initiate shorts. By the time the stock (or whatever instrument you're trading) is extended from its moving average it's time to cover. Then you can wait for another bounce to reload.

The strategy to cover & reload makes more sense when you're short due to the fact that if you're dead right with your call on the stock dropping the most you can gain is 100%. The odds of that happening are slim and if it ever did happen you'd likely have to ride out some severe short squeezes. But you might be able to stick & move your way to more that a 100% gain by reloading multiple times. In theory, you could catch 15 10% moves lower in a stock that's falling, retracing & falling anew.

Short sellers need to be nimble, pick their entry and exit spots wisely and heed signs of impending reversals. Trying to short breakdowns of an already extended market is a sucker's play -- as is overstaying your welcome while short.

I got this question the other day:

When you short a stock at 30.00 and you want to place a stop limit in case it jumps up past your max loss of say, 31.00, how exactly is that order entered (everytime I try to do this, it sells (Ed: he means buys) my stock immediately - I'm doing something really wrong). Do I put a limit price of 31.00 and a stop price above that, i.e. 31.50 (thus saying it's ok to repurchase any shares between 31.00 and 31.50)?

This is one of the trickier orders to place. In that situation you would want to enter a buy stop limit order (buy on stop with a limit). The stop would be at 31.00 and the limit would be 31.50. Here's a screen shot of a similar order to cover QQQQ with a buy stop @ 47 with a limit of 47.10:

What would happen here is that once the stop price of 47.00 is hit a buy limit order of 47.10 would be entered. As with any limit order, you're guaranteed price but not execution, so you may not be filled. If you wanted to be sure that you got out (covered) just use a buy stop of $47.00.

In case you missed it in the links section the other day, be sure to read the interview with Michelle B over on the StockTickr blog.

The other day Downtown Trader asked me how I handled trailing my stop on that DGX short. I pretty much followed exactly what's in my post about how I take partial profits. I moved the stop to breakeven once my gain was equal to my initial risk of 50 cents per share. As it dropped further I moved the stop to lock in one-third of my profit. Why one-third? It's a value I came up with which I think allows a stock to have a normal retracement without stopping me out. It's sort of an arbitrary level, although I got the idea from the Fibonacci levels. I've found that method to work reasonably well after much trial and error with trailing stops. I read somewhere that you can consider a retracement to be normal as long as it didn't break the 62% Fib level. So I figured that I could trail my stops somewhere around that level to allow stocks to have their normal fluctuations. I generally get more aggressive with the trailing stops as the day wears on. I'll move them to 50% by 3:00 and just keep pushing them until the close -- assuming I don't bail before then like I did with DGX.

The other day I got the following email about narrow range bars:

I was going through your site and others and I keep on hearing, no entry on narrow range bars. What is a narrow range bar, and could you send me a picture of one?

A narrow range bar (candle) is simply a bar which has a range from high to low that's much less than the average bar for a given equity. You'll often see people, like Dr. Sen, tracking NR7 bars. NR7 means the narrowest range of the last seven bars. For example, here are some NR7s from last Friday:

So now that you know what narrow range candles are, why should you care about them? Since stocks generally cycle between periods of low volatility and high volatility it can be beneficial to identify when the high volatility times are approaching. Often times, especially for trending charts, narrow range periods -- periods of range contraction -- will presage range expansion. You can think of a stocks movement like a spring / slinky -- it need to recoil (contract) in order to build up enough potential energy for its next expansion. However, it's important thing to note that range contraction doesn't tell you the direction of the impending expansion only that an expansion may be coming. It's up to the trader to use other means (indicators, common sense?) to catch the expansion in the right direction.

One of the nice things about trading against narrow range bars is that they give you a tight stop. If you size your positions like I do, the smaller the stop means you can buy (or short) more shares. It's the trades with the really tight stops that let you make those huge R-Multiple trades that often make or break a trader. Here's an example from a trade I took last week.

I'm going to use the actual entries taken by myself and by Butterboy but instead of using my actual exits, I'll assume I held until the close. So let's say my equity is $100,000 and my R, the amount I want to risk per trade, is 1% or $1,000. Given that, here's the first trade, based off of my actual entry parameters:

• Entry at 53.96
• Stop was 50 cents higher so I could short 1000/.50 or 2,000 shares
• DGX closed at 49.64, which is $4.32 lower. So that trade returned $8,640 or 8.64 times my risk (8.64R)

Here's the DGX trade using Butterboy's entry criteria. He used a 15 minute chart and entered lower than I did but his stop was also much tighter:

• Entry at 53.40
• I believe his stop was 21 cents higher, just above the 11:30 bar which he entered against. Given that one could short 1,000/.21 or just over 4,700 shares.
• DGX closed at $3.76 beneath the entry so that trade returned $17,672 or 17.67 times the initial risk (17.64R).

It should be clear why so many of us like narrow range bars and why I scan for them. That second trade had a worse (lower) entry but because of its tighter stop it was almost twice as profitable. It also shows why I love the percent risk position sizing model.

Here's an article on range contraction written by Oswald S. Castillo, TheStockStalker and very possibly the #1 fan of range contraction trading. Oswald also has a webinar on his site entitled "Principles of Trading Range Contraction". It's a good video but you have to register to view it.

You may also find Alan Farley's work on "coiled springs" and NR7s helpful. He goes in-depth on coiled springs in his book, The Master Swing Trader.

P.S. I almost forgot to answer the question of: How Narrow is Narrow?

For daytrading I like to find candles that are 1% or less than the price of the stock. So for a $20 stock any candles with a range of 20 cents or less really get my attention. In cases like those I just need small moves in the stock to rack up some nice gains. I'll typically pass on candles with a range much greater than 1% because they'll require a much larger move in the stock in order for me to get those large R-multiples.

For swing trading (looking at daily charts) I don't have a concrete answer on how narrow is narrow. But you could probably come up with some value based on the stock's average true range (ATR) or some other method. For example, you could define narrow as less than half a stocks ATR.

NRx (narrowest range on the last X candles) is always a good way to go as well, and it's easy to scan for.

Michelle B submits: I remember reading about a French housewife who when asked by her husband every morning what they would be having for dinner would always give the same tireless reply: "It depends on what jumps into my market basket." Some traders prefer to trade from a prepared watchlist. For me, trading is interesting because I do not know what will jump into my market basket. My preparation is my confidence in my ability to execute perceived opportunities according to my risk parameters.

Using the premarket top gainers/losers scan, I search the NASDAQ, NYSE, and AMEX market stalls for any tasty morsels and look for any stocks moving on the highs/lows scan. I also check Briefing.com and MarketWatch Newsfinder for news that has happened since the close of the previous trading day---identifying which stocks the market considers to be newsworthy is one of the many trading skills which can be developed.

In addition, I note via Briefing.com calendars such events as economic reports, earnings, Fedspeak, splits, upgrades/downgrades, and conferences. Usually, I have around ten to twelve trading candidates. Sometimes, there are none or a very small number. In that case, I patiently wait until the market has a new batch of fresh candidates---I will not make do with stale merchandise.

I list the stock symbols of my candidates and the important economic events for that day in my journal. Next to each stock symbol, I put the results of my basic Yahoo financial research---reason for price movement, float size, and short interest. I identify support and resistance on various timeframes. Sometimes, I check Yahoo Finance message boards and blogs to find out why a stock may be moving. Though in that case, I am always on my guard to weed the chaff from the wheat.

Then, I place my candidates on the chart pages of my Advanced Trading Platform (ATP). Four thirty-minute charts go on pages two through four. The main page contains the following windows: order entry, order status, real time portfolio, high/low scan, top gainers/losers scan, watchlist containing, in addition to my trading candidates, the major indices and technical indicators like VIX and TICK. Last but not least, is a chart window for the most promising candidate.

Discovering that ININ---set to gap up because it has given higher earnings' guidance---has a very small float and moderately high short interest, coupled with my knowledge that such stocks often move very strongly on low volume days, like Fridays, pre-holidays, and summer days, makes ININ the star attraction at this stage.

The market opens, and since I am particularly interested in seeing if ININ can trigger a trade, it gets top attention with a one-minute chart (above) on my main trading page. Meanwhile, I am flipping through my chart pages to see if anything is close to triggering---nothing yet. I check the one-minute chart of ININ, and I see the choppy pennant being formed at 10 A.M. The pennant range is around .25. That's enough of a sign for me to buy half of my lot as close to the bottom of this formation at 13.65. Buying a partial lot before the actual breakout happens makes sense in high-probability trades, especially if the desired lot size is substantial. I am set to buy the rest via an automatic buy stop if it can clear the top of the pennant at 13.87 on the 5 minute (see below). Trades moving strongly in the morning require monitoring via shorter time frames.

Fifteen minutes later, it does, and my full lot is in, average cost around 13.78. Just before the breakout, the pennant was around seventy-five percent completed and the volume dried up. Triangles/pennants need to be not completely formed or else they usually just fizzle out and do not support strong breakouts. The pause in volume is what my trading buddy calls a lullipop, a lull before the breakout, where the equilibrium between buyers and sellers end and the buyers take over.

Part of my prepurchase research of ININ was its support and resistance based on weekly price level resistance. Since the open of the July 24, 2006 weekly, bearish, high-volume candle was 14.81 (see below), therefore trapping lots of traders at that level, I determined the resistance to be around 14.80. The target of 14.80 also matches the length of the pole on which the flag formed on Friday's intraday charts, forming a measured move and confirming the resistance.

As the ININ trade continues, I am monitoring the price action on different timeframes, and have my mental stop under the flag for the first half lot at 13.53 and then just above the flag at 13.89 for the complete lot. Price is smoothly rising---each candle opens near but not with much overlapping to the close of the previous candle. I put my sell limit in for the full lot just under the target price. My offer gets lifted fairly easily, and I am out with a $1.00 move gotten in about 40 minutes. Since the risk was .25 for the complete lot, it was a +4R trade. For quickly moving stocks, I will often forgo hard stops, but it is a risk because my net connection could be interrupted. Then my carefully constructed trade will exist only in cyberspace floating without my pilotage. So, I am working on the artful science of consistently setting hard regular/trailing stops with upper triggers via bracketed orders to take care of that possibility.

My market basket is now empty until the next fresh thing jumps into it.

There seems to be a lot of controversial over the concept of R-Multiples. I've been seeing people complain about them for months now and I've been meaning to write a post about "R". I really wanted to do it last week but I'm glad I didn't get to it because this week a raging debate about R has popped up. Glenn, at DehTrader can serve as the poster child for the anti-R crew. Here's part of his recent rant against R-Multiples (emphasis is mine):

I post real numbers as opposed to R values, I always have. I like real numbers, I understand real numbers and I see truth in real numbers and I think the reader does too. As a reader of many blogs I find zero value in any post or summaries containing R values, I don't see any point in sharing that information. I suppose if I posted in R values I could look like a pretty good trader, but we all know I am a struggling trader. R can mean anything so why even bother with it... R stands for bullsh!t imo and that's my rant (that and ads haha). The best blogs out there post real numbers, Boogtser, JC (NYSE), the Kirkster all come to mind.

He's joined by folks like Paul who left this comment over on Ugly's post about R multiples:

I believe dollar values are more important than R value. I agree that the actual $ value is meaningless. However, R values are subjective and don’t give you a true idea on how successful the trade was. If you defined your risk at 15 cents and made 30 cents on the trade, while another person made 50 cents but decided his risk would be 50 cents, R values would say the guy who made 30 cents was more successful. I have a problem with that. It could very well be that the guy who only risked 15 cents is playing it too safe and his 2R gain was a bad trade.

So that gives you an idea of the anti-R sentiment. I'm going to explain why I think R-Multiples are so useful and why I use them in my trading and on this site.

What is R?

R is simply the dollar risk per trade. It's nothing but a reward-to-risk ratio. I first heard it called "R" in Van Tharp's book "Trade Your Way to Financial Freedom". In another of his books, "Financial Freedom Through Electronic Day Trading", Dr. Tharp reveals the great secret of trading:

The golden rule of trading is to keep losses at a level of 1 R as often as possible and to make profits that are high-R multiples.

You often hear (read) that traders should only look for trades with a reward/risk ratio of at least 2 or 3 to 1. Expressing your results in terms of how many times your risk allows you to easily see how well your trades measure up to such a standard. So when I look at my results in terms of multiples of R I can easily tell how good or bad the trades were. I like to think of R-Multiples as telling you the efficiency of your system.

So why not just use dollars?

Expressing my results in dollars would achieve the same result if I always risked the same amount of money. But what if I triple my account and therefore trade larger positions compared to when I started trading? Or what if I hit a rough spot and decide to cut my share size down while I ride out the storm? Then the dollar results won't easily tell me how trades from one period of time compared to another period of time. But if I use R making such comparisons is simple. Either my trades passed the risk / reward ratio test or they didn't. The actual number of dollars at risk doesn't matter, how many multiples of the dollars at risk does.

Along the same lines, recording trades in terms of R-Multiples allows you to easily calculate your system's expectancy. (Follow the link for why you should care about expectancy.)

Also, as Rx said:

talking and thinking in terms of R-multiple when you discuss about profits is an excellent approach - that by itself makes you focus on risk and money management - the actual "grail" to successful trading.

That is a very important point. Whenever I see people posting dollar returns, especially losses, that are all over the place the first thing I ask myself is "I wonder what his risk per trade is". It's almost a certainty that those traders aren't focusing on risk and as a result keep having huge losses. The mere fact that you have to define R and then place a stop to keep your loss to 1R is probably too constraining for those gamblers traders. Dr. Tharp says about determining your initial stop-loss point as soon as you enter a trade, which, by definition woud give you a 1 R loss:

This principle is so important that if you cannot follow it, then you might as well give up the idea of electronic day trading right away.

The reason I use R on the blog is because I don't want to discuss dollars or my account size on the site (as Ugly stated). That's nobody's business but mine. Also, it makes it easy for people to figure out what they could have made or lost on a trade with their own account size and risk per trade amount. If you see a trade that returned 3R all one has to do is plug in their dollar risk per trade to figure out what they could have made / lost.

To the R Haters

Let me address the "alleged" issues which I quoted above...

Glenn thinks that R is just some made up number and could mean anything. He likes "real" numbers. While it may be fun to see that somebody made $10,000 on a trade that in and of itself doesn't tell you how good that trade was. What if that person risked $30,000 to make that $10,000? Or what if they risked $1,000 to make that $10,000? Those are two very different trades. Sure they both made the same amount of money but isn't the second trade a much more efficient use of capital?

What if somebody is trading $500,000 lots to make $1,000 in profit? It may be nice to see somebody saying that they made $1,000 here and $1,000 there but damn(!) that's an inefficient use of capital. So while R could mean anything in terms of dollars, in my humble opinion what really matters is how many multiples of R were made or lost. That tells you the quality of a trade or system.

Glenn also states that if he reported his trades in terms of R he could appear to be a good trader. I'm sorry to tell him that's simply not the case. If you lost money that means your expectancy, which is just your average return expressed in R-Multiples, was negative.

Paul said that "R values are subjective and don’t give you a true idea on how successful the trade was". That is exactly wrong. R-multiples are the very thing that tells you exactly how successful a given trade was, if you choose to grade on a risk/reward basis.

Percentages vs. Dollars

This debate about R reminds me of a conversation I had a couple of weeks ago. I was in a presentation for Trade-Ideas' new tool, the Odds Maker. They were showing how you could backtest all these different scenarios with the tool. The results were expressed in average dollars won or lost. Another viewer and I asked about seeing the results in percentages. They kept saying that perhaps they would do that in a later revision. I kept harping on it because to me seeing the results in dollars was of little use for the way I size my positions

The argument from the presenter was that all you had to do was multiply the average dollar return by your average lot size to figure out how much money you could have made with a given system you were testing. I had to disagree because my lot size can vary drastically depending on how far away my stop loss is. Here's a situation which could be problematic -- I trade Google with a 2 point stop (which is only about half of a percent) and get lucky and make 6 points of profit. All of my other trades are on stocks under $50 with stops less than 50 cents. I could have some combination of winners and losers mixed in there... most of them probably well less than $6. That $6 gain may skew the results when presented as average dollars won. That's an over-simplification and there are all kinds of possible permutations. But I hope my point is clear that looking at the results in terms of average dollars won/lost may not tell accurately tell you the story.

So how can we make the results clearer? Simple, express them in percentages. That way, regardless of how many shares were traded or the prices of the stocks traded the results can be equalized across all the trades. I feel much better being able to say , "OK, this system would have returned X%" instead of "X number of points.

We debated the merits of each way of reporting for a few minutes and at one point somebody said, well , for this release we're aiming for the "lowest common denominator". In other words, the average person can't think in percentages, so we're just gonna report in points. I was like, F the average person, make it work the "right" way! The funny thing is that after debating all of that the software actually could express the results in percentage terms. We just had to switch a setting.

So my point of that little story is that I always prefer to think in terms of percentages in stead of points. I always see people talking about number of shares of point moves. For example, you might hear somebody exclaim "Google is up 5 points!!!" I don't see that as anything to get excited about. That just over a 1% move -- a normal fluctuation. You'll hear similar things from reporters talking excitedly about the Dow being up some triple-digit amount. The Nasdaq may actually be up a lot more on a percentage basis but they'll just say, eh, the Nasdaq is "just" up 30 points.

Looking at the percentages makes those kind of comparisons easier. R-Multiples do the same thing for traders. They can accurately compare their own trades and they can take another trader's results expressed in R and easily relate them to their own system.

Cameron just left a question for me in the comments to last night's recap. He asked:

I'm looking at your way of thinking about the market. One of the biggest questions I have, though, is regarding your stops.

You said you set your stop at just .20 over your short dollar amount. With all of the intra-day fluctuations, how can you know that it wouldn't have gone up, say, .25 and then nose-dived? It's a perfectly plausible situation.

Said another way, I can place a buy order with a stop, but we easily have 1% swings in a day. If a swing occurs, I have automatically sold for a lower price, only to see the market go back up. Therefore, I'm encouraged to set much broader stops. You set yours for only .20 over the buy price, though, which is why I'm confused.

I often get asked variations of this question, so I thought I'd put my answer in its own post. This was my response:

I couldn't have known that the stock wouldn't have shaken me out. But I also wouldn't have known if I'd have been shaken out if my stop was 50 cents or even a dollar higher. That narrow range candle just provided me with a tight stop. I like tight stops b/c they give me a better opportunity to get high R-multiple winners.

But to your point, how tight of a stop is too tight and will practically guarantee being stopped out? It's hard to answer that question but I generally shy away from any stop that's less than 0.5% of the stock price. I normally like my stops to be about 1% of the stock price. In general, I'm more comfortable with somewhat tighter stops on the QQQQ b/c of how liquid it is and b/c it's less volatile than an "in play" stock would be.

Also, note that I didn't pick a 20 cent stop out of thin air. I went with what the previous candle told me was resistance.

The key thing to remember is that I'm on the hunt for high R-Multiple winners. Because of the way I size my positions (the percent risk model), the tighter my stops the higher the potential reward. In other words if I think a stock has potential to move $2 I'd rather have a 20 cent stop than a $1 stop. Risking 20 cents to make 200 cents ($2) would give me a 10R winner vs only a 2R winner if I had risked $1. (If R was 1% of my equity that would be a 10% gain vs a 2% gain.) For a real example of this see the DGX trades I posted on my Narrow Range Bars post.

So there' a trade-off at work here between stop size and risk/reward multiples. The tighter the stop, the better chance you have of getting a high R winner but you also have a better chance of getting stopped out. Your win ratio is likely to be inversely proportional to how tight your stops are -- looser stops should lead to a higher win rate but also fewer high R-multiple winners.

It's also key that you don't pick your stops willy-nilly. I always use the support/resistance levels that the charts dictate to me.

There's some good advice in Alan Farley's latest article to help "home gamers" (Cramer's phrase) survive the current extra-volatile market. The whole article is worth reading but here are the four bits of advice:

I have four pieces of cautionary advice for my at-home brethren. Listen up, because it's no longer a question of whether you want to take real money out of the market, or just add a few bucks to the weekly shopping budget. These remedial steps must be taken if you want to survive long enough to take advantage of the real opportunity.

1. Wait for the Market Volatility Index (VIX) to drop below 40 and stay there for a week. Massive price swings require equally massive stop losses, which rarely justify the intended positions. Your only alternative is to stand aside and do nothing, no matter how much it hurts to watch others playing those big rallies and selloffs.
2. Forget overnight positions until the index futures stop gapping 2% or more every morning. These price jolts are great news when you're on the right side of the trade, but total devastation if you're on the wrong side. And guess what? You're not smart enough to predict overnight direction from day to day. Neither am I.
3. End your love affair with popular stocks that made you money during the last bull market. In November 2008, these are the issues that will trigger the most painful and unexpected reversals, which happen right after you're absolutely convinced your position is the right play. The bottom line: They see you coming, sucker.
4. Get control of the time element in your market strategy. You're getting killed because you have no patience and forgot how to sit on your hands when your trading edge isn't in play. Realistically, it could be months before the market works for you again. Would you rather wait it out and survive, or stay busy and get crushed?

My thoughts:

1. I like that part about watching the VIX. A few weeks ago I decided to cut my risk per trade (R) in half (from 1% to 0.5% of my equity) until the VIX breaks that top it's trying to put in. That means a close below 44. (Dr. Brett recently wrote about when traders should cut their risk.)
2. Tough to argue with what he said about overnight holds. This is certainly a market for intraday trades.
3. Reminds me of what I said about AAPL and GOOG back in July. Add DRYS and its ilk to the list too. It's better to just have an open mind and grab on to whichever stocks become the new leaders.
4. Yep, preservation of capital must be the top priority. If you can't sit out at least lower your risk.

I was recently asked about how I determine my "universe" of stocks. That is the stocks that I consider fair game to be traded. I use different filters for swing trading and day trading but they do have similar criteria. In both cases I'm trying to find liquid, volatile stocks. Once the filter's applied I then look for certain patterns / setups to trade.

For swing trade candidates I use the following criteria in TeleChart:

• Price Per Share: about $5 to Max
• 5 Day Average Volume: 75th percentile & higher (right now that works out to about 466,000 shares per day)
• Price Volatility: 25th percentile & higher
• Volume (Dollars) 5-Day: 75th percentile & higher

Here's a shot of the filter. The count column shows how many stocks are left after applying each row in the filter.

That typically narrows the market down to about 1,000 stocks. Then I look for setups among that set of stocks. I gave a lot of details about the setups I scan for back in the day.

For day trading I'm trying to find liquid stocks which are abnormally active and volatile. To do that I use the following two scans in Trade-Ideas:

Bearish Scanner's Filter:

• Min Price: $10
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: -0.5% (negative means down at least 0.5%)

Bullish Scanner's Filter:

• Min Price: $5
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: 1%

Sometimes I'll tweak these settings to give more or fewer alerts. I usually do that by adjusting the "Min Current Volume". Details about the actual alerts & setups I track for daytrading are on my "how I work" page. If you use Trade-Ideas you can paste the following link into the "Collaborate" feature and load that filter and all the alerts I track:

Bearish -- grab the URL of this link

Bullish -- grab the URL of this link

One thing I have always preached on the blog, less so on Twitter, where I bang out more trading ideas and market thoughts is that shorting stocks is hard. I think it’s harder than any aspect of learning the market. It’s dangerous. Let this morning be the only reminder you should EVER need.

Howard's right about shorting being difficult. I think it's so hard because it's not simply the opposite of going long. (I won't even go into the whole thing about your losses while short being theoretically infinite. Been there... done that. Use a stop to limit your losses, use proper position sizing, ONLY short LIQUID stocks and you'll be fine.) What makes shorting tricky is that bear moves often have violent (short-covering) rallies because the psychology of the crowd trading a down market is different than that of a bull market. You have to be quick on your feet when shorting. My motto is "stick & move".

Many traders love to buy breakouts in bull markets. (Whether that's actually a good strategy is debatable). In my experience swing trading, the opposite of that strategy, shorting breakdowns through resistance, will often lead you right into a snapback rally and, as MaoXian used to say "the quickest loss ever". That's why I often make note of all the people who are initiating shorts after the market has already fallen to a major support level. We saw that this week as I noted in some of the morning watchlists.

My contention is that if you were caught short Friday morning you should consider your losses as tuition paid to the school of hard knocks. Learn from that expensive lesson, take your losses and hopefully survive to trade another day. I'll stop short of saying that the losses were deserved but there were plenty of warnings to at least cover your shorts if not to get long. There are so many good sources of market information these days, both on the web and, yes, even on TV. I'm not saying to blindly follow somebody else's opinion but it can be helpful to see what others are doing based on what they see. Here are just a few of the recent warning signs:

• T2108 dropping below 20 -- Ah, good old reliable T2108. I've been watching it closely as we've sold off. On Wednesday I noted that it finally hit the point where wise shorts would want to cover. It pays to find a good overbought/oversold indicator and heed its warnings. (You may need different indicators or settings for different timeframes.) Sell at overbought and cover at oversold. A couple of years ago I decided to force myself to put some IRA money to work every time T2108 broke 20. It hasn't failed me yet.
• The Fed (Plunge Protection Team) has interfered announced stimulus packages around the July lows a couple times. On Thursday and earlier in the week there was talk of more PPT action.
• The PPT took action earlier in the week and last week. Just look at the orchestration of the LEH and AIG situations, etc.
• Extreme Volatility -- The moves all week were nothing short of violent. Positions were whipsawed all over the place. That in & of itself would be reason enough to lighten up on positions if not move to the sidelines. That kind of volatility is often a sign of a trend reversal, not of a continuation of the previous trend. That's why so many traders watch the VIX. Tons of people noted the spike in the VIX on Thursday.
• Corey from the 'Afraid to Trade' blog warned 'Use Extreme Caution in the Week Ahead.' He gave many good reasons, including the Federal Reserve interest rate decision, the quadruple witching options expiration and headline risk from troubled financial firms. His crystal ball was working well when he wrote "We could see a week ahead that will be discussed years later - as such, if you are a newer trader, it might be best to switch to simulation mode this week or use this week as a training experience, rather than risking real capital in an environment that could swing violently up and down due to market events scheduled to happen this week."
• Bullish Technical Divergences -- Dr. Brett pointed out some bullish divergences he was seeing in the market. Perhaps most important were what he called 'those fuzzy indicators' -- "Traffic on the blog is way up, reflecting trader uncertainty and desire for information. I just fielded my fourth media interview request in two days. During quiet and bullish market periods, I don't get four requests in a month." On Tuesday morning I also noted my blog's traffic spiked as did Barry Ritholtz. I've often joked about making some kind of sentiment indicator based on my site's traffic ebb & flow & referral logs. It's not a bad idea and I think it would be especially useful if it were based on a major financial site's traffic data.
• Dennis Gartman telling folks to "be small" -- Gartman was on CNBC's 'Fast Money' early in the week saying that he was scared of the market's movement and he was "being small" and planned to "get smaller". His advice to others was to "be small" in this market.
• Jeff Macke, also on 'Fast Money' was warning people not to play (trade) if they didn't understand the (changing) rules of the game. He was referring to all the headline risk from PPT action and the volatility caused by rumors -- many of which were spread by CNBC during the trading sessions. Ironically, Macke was kicking himself Thursday night for not following his own advice and getting caught short.

I fully believe that had the PPT not acted on Thursday night the market was *destined* to move higher in the short term on its own. Still being short on expiration Friday in this environment was just asking for trouble. So what's a trader to do? Like I said earlier, stick & move. I think it makes much more sense to short bounces back to a trendline or moving average. William O'Neil's book on short selling talks about using retracements to the 50 and/or 200-day moving average as a place to initiate shorts. By the time the stock (or whatever instrument you're trading) is extended from its moving average it's time to cover. Then you can wait for another bounce to reload.

The strategy to cover & reload makes more sense when you're short due to the fact that if you're dead right with your call on the stock dropping the most you can gain is 100%. The odds of that happening are slim and if it ever did happen you'd likely have to ride out some severe short squeezes. But you might be able to stick & move your way to more that a 100% gain by reloading multiple times. In theory, you could catch 15 10% moves lower in a stock that's falling, retracing & falling anew.

Short sellers need to be nimble, pick their entry and exit spots wisely and heed signs of impending reversals. Trying to short breakdowns of an already extended market is a sucker's play -- as is overstaying your welcome while short.

I got this question the other day:

When you short a stock at 30.00 and you want to place a stop limit in case it jumps up past your max loss of say, 31.00, how exactly is that order entered (everytime I try to do this, it sells (Ed: he means buys) my stock immediately - I'm doing something really wrong). Do I put a limit price of 31.00 and a stop price above that, i.e. 31.50 (thus saying it's ok to repurchase any shares between 31.00 and 31.50)?

This is one of the trickier orders to place. In that situation you would want to enter a buy stop limit order (buy on stop with a limit). The stop would be at 31.00 and the limit would be 31.50. Here's a screen shot of a similar order to cover QQQQ with a buy stop @ 47 with a limit of 47.10:

What would happen here is that once the stop price of 47.00 is hit a buy limit order of 47.10 would be entered. As with any limit order, you're guaranteed price but not execution, so you may not be filled. If you wanted to be sure that you got out (covered) just use a buy stop of $47.00.

In case you missed it in the links section the other day, be sure to read the interview with Michelle B over on the StockTickr blog.

The other day Downtown Trader asked me how I handled trailing my stop on that DGX short. I pretty much followed exactly what's in my post about how I take partial profits. I moved the stop to breakeven once my gain was equal to my initial risk of 50 cents per share. As it dropped further I moved the stop to lock in one-third of my profit. Why one-third? It's a value I came up with which I think allows a stock to have a normal retracement without stopping me out. It's sort of an arbitrary level, although I got the idea from the Fibonacci levels. I've found that method to work reasonably well after much trial and error with trailing stops. I read somewhere that you can consider a retracement to be normal as long as it didn't break the 62% Fib level. So I figured that I could trail my stops somewhere around that level to allow stocks to have their normal fluctuations. I generally get more aggressive with the trailing stops as the day wears on. I'll move them to 50% by 3:00 and just keep pushing them until the close -- assuming I don't bail before then like I did with DGX.

The other day I got the following email about narrow range bars:

I was going through your site and others and I keep on hearing, no entry on narrow range bars. What is a narrow range bar, and could you send me a picture of one?

A narrow range bar (candle) is simply a bar which has a range from high to low that's much less than the average bar for a given equity. You'll often see people, like Dr. Sen, tracking NR7 bars. NR7 means the narrowest range of the last seven bars. For example, here are some NR7s from last Friday:

So now that you know what narrow range candles are, why should you care about them? Since stocks generally cycle between periods of low volatility and high volatility it can be beneficial to identify when the high volatility times are approaching. Often times, especially for trending charts, narrow range periods -- periods of range contraction -- will presage range expansion. You can think of a stocks movement like a spring / slinky -- it need to recoil (contract) in order to build up enough potential energy for its next expansion. However, it's important thing to note that range contraction doesn't tell you the direction of the impending expansion only that an expansion may be coming. It's up to the trader to use other means (indicators, common sense?) to catch the expansion in the right direction.

One of the nice things about trading against narrow range bars is that they give you a tight stop. If you size your positions like I do, the smaller the stop means you can buy (or short) more shares. It's the trades with the really tight stops that let you make those huge R-Multiple trades that often make or break a trader. Here's an example from a trade I took last week.

I'm going to use the actual entries taken by myself and by Butterboy but instead of using my actual exits, I'll assume I held until the close. So let's say my equity is $100,000 and my R, the amount I want to risk per trade, is 1% or $1,000. Given that, here's the first trade, based off of my actual entry parameters:

• Entry at 53.96
• Stop was 50 cents higher so I could short 1000/.50 or 2,000 shares
• DGX closed at 49.64, which is $4.32 lower. So that trade returned $8,640 or 8.64 times my risk (8.64R)

Here's the DGX trade using Butterboy's entry criteria. He used a 15 minute chart and entered lower than I did but his stop was also much tighter:

• Entry at 53.40
• I believe his stop was 21 cents higher, just above the 11:30 bar which he entered against. Given that one could short 1,000/.21 or just over 4,700 shares.
• DGX closed at $3.76 beneath the entry so that trade returned $17,672 or 17.67 times the initial risk (17.64R).

It should be clear why so many of us like narrow range bars and why I scan for them. That second trade had a worse (lower) entry but because of its tighter stop it was almost twice as profitable. It also shows why I love the percent risk position sizing model.

Here's an article on range contraction written by Oswald S. Castillo, TheStockStalker and very possibly the #1 fan of range contraction trading. Oswald also has a webinar on his site entitled "Principles of Trading Range Contraction". It's a good video but you have to register to view it.

You may also find Alan Farley's work on "coiled springs" and NR7s helpful. He goes in-depth on coiled springs in his book, The Master Swing Trader.

P.S. I almost forgot to answer the question of: How Narrow is Narrow?

For daytrading I like to find candles that are 1% or less than the price of the stock. So for a $20 stock any candles with a range of 20 cents or less really get my attention. In cases like those I just need small moves in the stock to rack up some nice gains. I'll typically pass on candles with a range much greater than 1% because they'll require a much larger move in the stock in order for me to get those large R-multiples.

For swing trading (looking at daily charts) I don't have a concrete answer on how narrow is narrow. But you could probably come up with some value based on the stock's average true range (ATR) or some other method. For example, you could define narrow as less than half a stocks ATR.

NRx (narrowest range on the last X candles) is always a good way to go as well, and it's easy to scan for.

Michelle B submits: I remember reading about a French housewife who when asked by her husband every morning what they would be having for dinner would always give the same tireless reply: "It depends on what jumps into my market basket." Some traders prefer to trade from a prepared watchlist. For me, trading is interesting because I do not know what will jump into my market basket. My preparation is my confidence in my ability to execute perceived opportunities according to my risk parameters.

Using the premarket top gainers/losers scan, I search the NASDAQ, NYSE, and AMEX market stalls for any tasty morsels and look for any stocks moving on the highs/lows scan. I also check Briefing.com and MarketWatch Newsfinder for news that has happened since the close of the previous trading day---identifying which stocks the market considers to be newsworthy is one of the many trading skills which can be developed.

In addition, I note via Briefing.com calendars such events as economic reports, earnings, Fedspeak, splits, upgrades/downgrades, and conferences. Usually, I have around ten to twelve trading candidates. Sometimes, there are none or a very small number. In that case, I patiently wait until the market has a new batch of fresh candidates---I will not make do with stale merchandise.

I list the stock symbols of my candidates and the important economic events for that day in my journal. Next to each stock symbol, I put the results of my basic Yahoo financial research---reason for price movement, float size, and short interest. I identify support and resistance on various timeframes. Sometimes, I check Yahoo Finance message boards and blogs to find out why a stock may be moving. Though in that case, I am always on my guard to weed the chaff from the wheat.

Then, I place my candidates on the chart pages of my Advanced Trading Platform (ATP). Four thirty-minute charts go on pages two through four. The main page contains the following windows: order entry, order status, real time portfolio, high/low scan, top gainers/losers scan, watchlist containing, in addition to my trading candidates, the major indices and technical indicators like VIX and TICK. Last but not least, is a chart window for the most promising candidate.

Discovering that ININ---set to gap up because it has given higher earnings' guidance---has a very small float and moderately high short interest, coupled with my knowledge that such stocks often move very strongly on low volume days, like Fridays, pre-holidays, and summer days, makes ININ the star attraction at this stage.

The market opens, and since I am particularly interested in seeing if ININ can trigger a trade, it gets top attention with a one-minute chart (above) on my main trading page. Meanwhile, I am flipping through my chart pages to see if anything is close to triggering---nothing yet. I check the one-minute chart of ININ, and I see the choppy pennant being formed at 10 A.M. The pennant range is around .25. That's enough of a sign for me to buy half of my lot as close to the bottom of this formation at 13.65. Buying a partial lot before the actual breakout happens makes sense in high-probability trades, especially if the desired lot size is substantial. I am set to buy the rest via an automatic buy stop if it can clear the top of the pennant at 13.87 on the 5 minute (see below). Trades moving strongly in the morning require monitoring via shorter time frames.

Fifteen minutes later, it does, and my full lot is in, average cost around 13.78. Just before the breakout, the pennant was around seventy-five percent completed and the volume dried up. Triangles/pennants need to be not completely formed or else they usually just fizzle out and do not support strong breakouts. The pause in volume is what my trading buddy calls a lullipop, a lull before the breakout, where the equilibrium between buyers and sellers end and the buyers take over.

Part of my prepurchase research of ININ was its support and resistance based on weekly price level resistance. Since the open of the July 24, 2006 weekly, bearish, high-volume candle was 14.81 (see below), therefore trapping lots of traders at that level, I determined the resistance to be around 14.80. The target of 14.80 also matches the length of the pole on which the flag formed on Friday's intraday charts, forming a measured move and confirming the resistance.

As the ININ trade continues, I am monitoring the price action on different timeframes, and have my mental stop under the flag for the first half lot at 13.53 and then just above the flag at 13.89 for the complete lot. Price is smoothly rising---each candle opens near but not with much overlapping to the close of the previous candle. I put my sell limit in for the full lot just under the target price. My offer gets lifted fairly easily, and I am out with a $1.00 move gotten in about 40 minutes. Since the risk was .25 for the complete lot, it was a +4R trade. For quickly moving stocks, I will often forgo hard stops, but it is a risk because my net connection could be interrupted. Then my carefully constructed trade will exist only in cyberspace floating without my pilotage. So, I am working on the artful science of consistently setting hard regular/trailing stops with upper triggers via bracketed orders to take care of that possibility.

My market basket is now empty until the next fresh thing jumps into it.

There seems to be a lot of controversial over the concept of R-Multiples. I've been seeing people complain about them for months now and I've been meaning to write a post about "R". I really wanted to do it last week but I'm glad I didn't get to it because this week a raging debate about R has popped up. Glenn, at DehTrader can serve as the poster child for the anti-R crew. Here's part of his recent rant against R-Multiples (emphasis is mine):

I post real numbers as opposed to R values, I always have. I like real numbers, I understand real numbers and I see truth in real numbers and I think the reader does too. As a reader of many blogs I find zero value in any post or summaries containing R values, I don't see any point in sharing that information. I suppose if I posted in R values I could look like a pretty good trader, but we all know I am a struggling trader. R can mean anything so why even bother with it... R stands for bullsh!t imo and that's my rant (that and ads haha). The best blogs out there post real numbers, Boogtser, JC (NYSE), the Kirkster all come to mind.

He's joined by folks like Paul who left this comment over on Ugly's post about R multiples:

I believe dollar values are more important than R value. I agree that the actual $ value is meaningless. However, R values are subjective and don’t give you a true idea on how successful the trade was. If you defined your risk at 15 cents and made 30 cents on the trade, while another person made 50 cents but decided his risk would be 50 cents, R values would say the guy who made 30 cents was more successful. I have a problem with that. It could very well be that the guy who only risked 15 cents is playing it too safe and his 2R gain was a bad trade.

So that gives you an idea of the anti-R sentiment. I'm going to explain why I think R-Multiples are so useful and why I use them in my trading and on this site.

What is R?

R is simply the dollar risk per trade. It's nothing but a reward-to-risk ratio. I first heard it called "R" in Van Tharp's book "Trade Your Way to Financial Freedom". In another of his books, "Financial Freedom Through Electronic Day Trading", Dr. Tharp reveals the great secret of trading:

The golden rule of trading is to keep losses at a level of 1 R as often as possible and to make profits that are high-R multiples.

You often hear (read) that traders should only look for trades with a reward/risk ratio of at least 2 or 3 to 1. Expressing your results in terms of how many times your risk allows you to easily see how well your trades measure up to such a standard. So when I look at my results in terms of multiples of R I can easily tell how good or bad the trades were. I like to think of R-Multiples as telling you the efficiency of your system.

So why not just use dollars?

Expressing my results in dollars would achieve the same result if I always risked the same amount of money. But what if I triple my account and therefore trade larger positions compared to when I started trading? Or what if I hit a rough spot and decide to cut my share size down while I ride out the storm? Then the dollar results won't easily tell me how trades from one period of time compared to another period of time. But if I use R making such comparisons is simple. Either my trades passed the risk / reward ratio test or they didn't. The actual number of dollars at risk doesn't matter, how many multiples of the dollars at risk does.

Along the same lines, recording trades in terms of R-Multiples allows you to easily calculate your system's expectancy. (Follow the link for why you should care about expectancy.)

Also, as Rx said:

talking and thinking in terms of R-multiple when you discuss about profits is an excellent approach - that by itself makes you focus on risk and money management - the actual "grail" to successful trading.

That is a very important point. Whenever I see people posting dollar returns, especially losses, that are all over the place the first thing I ask myself is "I wonder what his risk per trade is". It's almost a certainty that those traders aren't focusing on risk and as a result keep having huge losses. The mere fact that you have to define R and then place a stop to keep your loss to 1R is probably too constraining for those gamblers traders. Dr. Tharp says about determining your initial stop-loss point as soon as you enter a trade, which, by definition woud give you a 1 R loss:

This principle is so important that if you cannot follow it, then you might as well give up the idea of electronic day trading right away.

The reason I use R on the blog is because I don't want to discuss dollars or my account size on the site (as Ugly stated). That's nobody's business but mine. Also, it makes it easy for people to figure out what they could have made or lost on a trade with their own account size and risk per trade amount. If you see a trade that returned 3R all one has to do is plug in their dollar risk per trade to figure out what they could have made / lost.

To the R Haters

Let me address the "alleged" issues which I quoted above...

Glenn thinks that R is just some made up number and could mean anything. He likes "real" numbers. While it may be fun to see that somebody made $10,000 on a trade that in and of itself doesn't tell you how good that trade was. What if that person risked $30,000 to make that $10,000? Or what if they risked $1,000 to make that $10,000? Those are two very different trades. Sure they both made the same amount of money but isn't the second trade a much more efficient use of capital?

What if somebody is trading $500,000 lots to make $1,000 in profit? It may be nice to see somebody saying that they made $1,000 here and $1,000 there but damn(!) that's an inefficient use of capital. So while R could mean anything in terms of dollars, in my humble opinion what really matters is how many multiples of R were made or lost. That tells you the quality of a trade or system.

Glenn also states that if he reported his trades in terms of R he could appear to be a good trader. I'm sorry to tell him that's simply not the case. If you lost money that means your expectancy, which is just your average return expressed in R-Multiples, was negative.

Paul said that "R values are subjective and don’t give you a true idea on how successful the trade was". That is exactly wrong. R-multiples are the very thing that tells you exactly how successful a given trade was, if you choose to grade on a risk/reward basis.

Percentages vs. Dollars

This debate about R reminds me of a conversation I had a couple of weeks ago. I was in a presentation for Trade-Ideas' new tool, the Odds Maker. They were showing how you could backtest all these different scenarios with the tool. The results were expressed in average dollars won or lost. Another viewer and I asked about seeing the results in percentages. They kept saying that perhaps they would do that in a later revision. I kept harping on it because to me seeing the results in dollars was of little use for the way I size my positions

The argument from the presenter was that all you had to do was multiply the average dollar return by your average lot size to figure out how much money you could have made with a given system you were testing. I had to disagree because my lot size can vary drastically depending on how far away my stop loss is. Here's a situation which could be problematic -- I trade Google with a 2 point stop (which is only about half of a percent) and get lucky and make 6 points of profit. All of my other trades are on stocks under $50 with stops less than 50 cents. I could have some combination of winners and losers mixed in there... most of them probably well less than $6. That $6 gain may skew the results when presented as average dollars won. That's an over-simplification and there are all kinds of possible permutations. But I hope my point is clear that looking at the results in terms of average dollars won/lost may not tell accurately tell you the story.

So how can we make the results clearer? Simple, express them in percentages. That way, regardless of how many shares were traded or the prices of the stocks traded the results can be equalized across all the trades. I feel much better being able to say , "OK, this system would have returned X%" instead of "X number of points.

We debated the merits of each way of reporting for a few minutes and at one point somebody said, well , for this release we're aiming for the "lowest common denominator". In other words, the average person can't think in percentages, so we're just gonna report in points. I was like, F the average person, make it work the "right" way! The funny thing is that after debating all of that the software actually could express the results in percentage terms. We just had to switch a setting.

So my point of that little story is that I always prefer to think in terms of percentages in stead of points. I always see people talking about number of shares of point moves. For example, you might hear somebody exclaim "Google is up 5 points!!!" I don't see that as anything to get excited about. That just over a 1% move -- a normal fluctuation. You'll hear similar things from reporters talking excitedly about the Dow being up some triple-digit amount. The Nasdaq may actually be up a lot more on a percentage basis but they'll just say, eh, the Nasdaq is "just" up 30 points.

Looking at the percentages makes those kind of comparisons easier. R-Multiples do the same thing for traders. They can accurately compare their own trades and they can take another trader's results expressed in R and easily relate them to their own system.

Cameron just left a question for me in the comments to last night's recap. He asked:

I'm looking at your way of thinking about the market. One of the biggest questions I have, though, is regarding your stops.

You said you set your stop at just .20 over your short dollar amount. With all of the intra-day fluctuations, how can you know that it wouldn't have gone up, say, .25 and then nose-dived? It's a perfectly plausible situation.

Said another way, I can place a buy order with a stop, but we easily have 1% swings in a day. If a swing occurs, I have automatically sold for a lower price, only to see the market go back up. Therefore, I'm encouraged to set much broader stops. You set yours for only .20 over the buy price, though, which is why I'm confused.

I often get asked variations of this question, so I thought I'd put my answer in its own post. This was my response:

I couldn't have known that the stock wouldn't have shaken me out. But I also wouldn't have known if I'd have been shaken out if my stop was 50 cents or even a dollar higher. That narrow range candle just provided me with a tight stop. I like tight stops b/c they give me a better opportunity to get high R-multiple winners.

But to your point, how tight of a stop is too tight and will practically guarantee being stopped out? It's hard to answer that question but I generally shy away from any stop that's less than 0.5% of the stock price. I normally like my stops to be about 1% of the stock price. In general, I'm more comfortable with somewhat tighter stops on the QQQQ b/c of how liquid it is and b/c it's less volatile than an "in play" stock would be.

Also, note that I didn't pick a 20 cent stop out of thin air. I went with what the previous candle told me was resistance.

The key thing to remember is that I'm on the hunt for high R-Multiple winners. Because of the way I size my positions (the percent risk model), the tighter my stops the higher the potential reward. In other words if I think a stock has potential to move $2 I'd rather have a 20 cent stop than a $1 stop. Risking 20 cents to make 200 cents ($2) would give me a 10R winner vs only a 2R winner if I had risked $1. (If R was 1% of my equity that would be a 10% gain vs a 2% gain.) For a real example of this see the DGX trades I posted on my Narrow Range Bars post.

So there' a trade-off at work here between stop size and risk/reward multiples. The tighter the stop, the better chance you have of getting a high R winner but you also have a better chance of getting stopped out. Your win ratio is likely to be inversely proportional to how tight your stops are -- looser stops should lead to a higher win rate but also fewer high R-multiple winners.

It's also key that you don't pick your stops willy-nilly. I always use the support/resistance levels that the charts dictate to me.

There's some good advice in Alan Farley's latest article to help "home gamers" (Cramer's phrase) survive the current extra-volatile market. The whole article is worth reading but here are the four bits of advice:

I have four pieces of cautionary advice for my at-home brethren. Listen up, because it's no longer a question of whether you want to take real money out of the market, or just add a few bucks to the weekly shopping budget. These remedial steps must be taken if you want to survive long enough to take advantage of the real opportunity.

1. Wait for the Market Volatility Index (VIX) to drop below 40 and stay there for a week. Massive price swings require equally massive stop losses, which rarely justify the intended positions. Your only alternative is to stand aside and do nothing, no matter how much it hurts to watch others playing those big rallies and selloffs.
2. Forget overnight positions until the index futures stop gapping 2% or more every morning. These price jolts are great news when you're on the right side of the trade, but total devastation if you're on the wrong side. And guess what? You're not smart enough to predict overnight direction from day to day. Neither am I.
3. End your love affair with popular stocks that made you money during the last bull market. In November 2008, these are the issues that will trigger the most painful and unexpected reversals, which happen right after you're absolutely convinced your position is the right play. The bottom line: They see you coming, sucker.
4. Get control of the time element in your market strategy. You're getting killed because you have no patience and forgot how to sit on your hands when your trading edge isn't in play. Realistically, it could be months before the market works for you again. Would you rather wait it out and survive, or stay busy and get crushed?

My thoughts:

1. I like that part about watching the VIX. A few weeks ago I decided to cut my risk per trade (R) in half (from 1% to 0.5% of my equity) until the VIX breaks that top it's trying to put in. That means a close below 44. (Dr. Brett recently wrote about when traders should cut their risk.)
2. Tough to argue with what he said about overnight holds. This is certainly a market for intraday trades.
3. Reminds me of what I said about AAPL and GOOG back in July. Add DRYS and its ilk to the list too. It's better to just have an open mind and grab on to whichever stocks become the new leaders.
4. Yep, preservation of capital must be the top priority. If you can't sit out at least lower your risk.

I was recently asked about how I determine my "universe" of stocks. That is the stocks that I consider fair game to be traded. I use different filters for swing trading and day trading but they do have similar criteria. In both cases I'm trying to find liquid, volatile stocks. Once the filter's applied I then look for certain patterns / setups to trade.

For swing trade candidates I use the following criteria in TeleChart:

• Price Per Share: about $5 to Max
• 5 Day Average Volume: 75th percentile & higher (right now that works out to about 466,000 shares per day)
• Price Volatility: 25th percentile & higher
• Volume (Dollars) 5-Day: 75th percentile & higher

Here's a shot of the filter. The count column shows how many stocks are left after applying each row in the filter.

That typically narrows the market down to about 1,000 stocks. Then I look for setups among that set of stocks. I gave a lot of details about the setups I scan for back in the day.

For day trading I'm trying to find liquid stocks which are abnormally active and volatile. To do that I use the following two scans in Trade-Ideas:

Bearish Scanner's Filter:

• Min Price: $10
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: -0.5% (negative means down at least 0.5%)

Bullish Scanner's Filter:

• Min Price: $5
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: 1%

Sometimes I'll tweak these settings to give more or fewer alerts. I usually do that by adjusting the "Min Current Volume". Details about the actual alerts & setups I track for daytrading are on my "how I work" page. If you use Trade-Ideas you can paste the following link into the "Collaborate" feature and load that filter and all the alerts I track:

Bearish -- grab the URL of this link

Bullish -- grab the URL of this link

One thing I have always preached on the blog, less so on Twitter, where I bang out more trading ideas and market thoughts is that shorting stocks is hard. I think it’s harder than any aspect of learning the market. It’s dangerous. Let this morning be the only reminder you should EVER need.

Howard's right about shorting being difficult. I think it's so hard because it's not simply the opposite of going long. (I won't even go into the whole thing about your losses while short being theoretically infinite. Been there... done that. Use a stop to limit your losses, use proper position sizing, ONLY short LIQUID stocks and you'll be fine.) What makes shorting tricky is that bear moves often have violent (short-covering) rallies because the psychology of the crowd trading a down market is different than that of a bull market. You have to be quick on your feet when shorting. My motto is "stick & move".

Many traders love to buy breakouts in bull markets. (Whether that's actually a good strategy is debatable). In my experience swing trading, the opposite of that strategy, shorting breakdowns through resistance, will often lead you right into a snapback rally and, as MaoXian used to say "the quickest loss ever". That's why I often make note of all the people who are initiating shorts after the market has already fallen to a major support level. We saw that this week as I noted in some of the morning watchlists.

My contention is that if you were caught short Friday morning you should consider your losses as tuition paid to the school of hard knocks. Learn from that expensive lesson, take your losses and hopefully survive to trade another day. I'll stop short of saying that the losses were deserved but there were plenty of warnings to at least cover your shorts if not to get long. There are so many good sources of market information these days, both on the web and, yes, even on TV. I'm not saying to blindly follow somebody else's opinion but it can be helpful to see what others are doing based on what they see. Here are just a few of the recent warning signs:

• T2108 dropping below 20 -- Ah, good old reliable T2108. I've been watching it closely as we've sold off. On Wednesday I noted that it finally hit the point where wise shorts would want to cover. It pays to find a good overbought/oversold indicator and heed its warnings. (You may need different indicators or settings for different timeframes.) Sell at overbought and cover at oversold. A couple of years ago I decided to force myself to put some IRA money to work every time T2108 broke 20. It hasn't failed me yet.
• The Fed (Plunge Protection Team) has interfered announced stimulus packages around the July lows a couple times. On Thursday and earlier in the week there was talk of more PPT action.
• The PPT took action earlier in the week and last week. Just look at the orchestration of the LEH and AIG situations, etc.
• Extreme Volatility -- The moves all week were nothing short of violent. Positions were whipsawed all over the place. That in & of itself would be reason enough to lighten up on positions if not move to the sidelines. That kind of volatility is often a sign of a trend reversal, not of a continuation of the previous trend. That's why so many traders watch the VIX. Tons of people noted the spike in the VIX on Thursday.
• Corey from the 'Afraid to Trade' blog warned 'Use Extreme Caution in the Week Ahead.' He gave many good reasons, including the Federal Reserve interest rate decision, the quadruple witching options expiration and headline risk from troubled financial firms. His crystal ball was working well when he wrote "We could see a week ahead that will be discussed years later - as such, if you are a newer trader, it might be best to switch to simulation mode this week or use this week as a training experience, rather than risking real capital in an environment that could swing violently up and down due to market events scheduled to happen this week."
• Bullish Technical Divergences -- Dr. Brett pointed out some bullish divergences he was seeing in the market. Perhaps most important were what he called 'those fuzzy indicators' -- "Traffic on the blog is way up, reflecting trader uncertainty and desire for information. I just fielded my fourth media interview request in two days. During quiet and bullish market periods, I don't get four requests in a month." On Tuesday morning I also noted my blog's traffic spiked as did Barry Ritholtz. I've often joked about making some kind of sentiment indicator based on my site's traffic ebb & flow & referral logs. It's not a bad idea and I think it would be especially useful if it were based on a major financial site's traffic data.
• Dennis Gartman telling folks to "be small" -- Gartman was on CNBC's 'Fast Money' early in the week saying that he was scared of the market's movement and he was "being small" and planned to "get smaller". His advice to others was to "be small" in this market.
• Jeff Macke, also on 'Fast Money' was warning people not to play (trade) if they didn't understand the (changing) rules of the game. He was referring to all the headline risk from PPT action and the volatility caused by rumors -- many of which were spread by CNBC during the trading sessions. Ironically, Macke was kicking himself Thursday night for not following his own advice and getting caught short.

I fully believe that had the PPT not acted on Thursday night the market was *destined* to move higher in the short term on its own. Still being short on expiration Friday in this environment was just asking for trouble. So what's a trader to do? Like I said earlier, stick & move. I think it makes much more sense to short bounces back to a trendline or moving average. William O'Neil's book on short selling talks about using retracements to the 50 and/or 200-day moving average as a place to initiate shorts. By the time the stock (or whatever instrument you're trading) is extended from its moving average it's time to cover. Then you can wait for another bounce to reload.

The strategy to cover & reload makes more sense when you're short due to the fact that if you're dead right with your call on the stock dropping the most you can gain is 100%. The odds of that happening are slim and if it ever did happen you'd likely have to ride out some severe short squeezes. But you might be able to stick & move your way to more that a 100% gain by reloading multiple times. In theory, you could catch 15 10% moves lower in a stock that's falling, retracing & falling anew.

Short sellers need to be nimble, pick their entry and exit spots wisely and heed signs of impending reversals. Trying to short breakdowns of an already extended market is a sucker's play -- as is overstaying your welcome while short.

I got this question the other day:

When you short a stock at 30.00 and you want to place a stop limit in case it jumps up past your max loss of say, 31.00, how exactly is that order entered (everytime I try to do this, it sells (Ed: he means buys) my stock immediately - I'm doing something really wrong). Do I put a limit price of 31.00 and a stop price above that, i.e. 31.50 (thus saying it's ok to repurchase any shares between 31.00 and 31.50)?

This is one of the trickier orders to place. In that situation you would want to enter a buy stop limit order (buy on stop with a limit). The stop would be at 31.00 and the limit would be 31.50. Here's a screen shot of a similar order to cover QQQQ with a buy stop @ 47 with a limit of 47.10:

What would happen here is that once the stop price of 47.00 is hit a buy limit order of 47.10 would be entered. As with any limit order, you're guaranteed price but not execution, so you may not be filled. If you wanted to be sure that you got out (covered) just use a buy stop of $47.00.

In case you missed it in the links section the other day, be sure to read the interview with Michelle B over on the StockTickr blog.

The other day Downtown Trader asked me how I handled trailing my stop on that DGX short. I pretty much followed exactly what's in my post about how I take partial profits. I moved the stop to breakeven once my gain was equal to my initial risk of 50 cents per share. As it dropped further I moved the stop to lock in one-third of my profit. Why one-third? It's a value I came up with which I think allows a stock to have a normal retracement without stopping me out. It's sort of an arbitrary level, although I got the idea from the Fibonacci levels. I've found that method to work reasonably well after much trial and error with trailing stops. I read somewhere that you can consider a retracement to be normal as long as it didn't break the 62% Fib level. So I figured that I could trail my stops somewhere around that level to allow stocks to have their normal fluctuations. I generally get more aggressive with the trailing stops as the day wears on. I'll move them to 50% by 3:00 and just keep pushing them until the close -- assuming I don't bail before then like I did with DGX.

The other day I got the following email about narrow range bars:

I was going through your site and others and I keep on hearing, no entry on narrow range bars. What is a narrow range bar, and could you send me a picture of one?

A narrow range bar (candle) is simply a bar which has a range from high to low that's much less than the average bar for a given equity. You'll often see people, like Dr. Sen, tracking NR7 bars. NR7 means the narrowest range of the last seven bars. For example, here are some NR7s from last Friday:

So now that you know what narrow range candles are, why should you care about them? Since stocks generally cycle between periods of low volatility and high volatility it can be beneficial to identify when the high volatility times are approaching. Often times, especially for trending charts, narrow range periods -- periods of range contraction -- will presage range expansion. You can think of a stocks movement like a spring / slinky -- it need to recoil (contract) in order to build up enough potential energy for its next expansion. However, it's important thing to note that range contraction doesn't tell you the direction of the impending expansion only that an expansion may be coming. It's up to the trader to use other means (indicators, common sense?) to catch the expansion in the right direction.

One of the nice things about trading against narrow range bars is that they give you a tight stop. If you size your positions like I do, the smaller the stop means you can buy (or short) more shares. It's the trades with the really tight stops that let you make those huge R-Multiple trades that often make or break a trader. Here's an example from a trade I took last week.

I'm going to use the actual entries taken by myself and by Butterboy but instead of using my actual exits, I'll assume I held until the close. So let's say my equity is $100,000 and my R, the amount I want to risk per trade, is 1% or $1,000. Given that, here's the first trade, based off of my actual entry parameters:

• Entry at 53.96
• Stop was 50 cents higher so I could short 1000/.50 or 2,000 shares
• DGX closed at 49.64, which is $4.32 lower. So that trade returned $8,640 or 8.64 times my risk (8.64R)

Here's the DGX trade using Butterboy's entry criteria. He used a 15 minute chart and entered lower than I did but his stop was also much tighter:

• Entry at 53.40
• I believe his stop was 21 cents higher, just above the 11:30 bar which he entered against. Given that one could short 1,000/.21 or just over 4,700 shares.
• DGX closed at $3.76 beneath the entry so that trade returned $17,672 or 17.67 times the initial risk (17.64R).

It should be clear why so many of us like narrow range bars and why I scan for them. That second trade had a worse (lower) entry but because of its tighter stop it was almost twice as profitable. It also shows why I love the percent risk position sizing model.

Here's an article on range contraction written by Oswald S. Castillo, TheStockStalker and very possibly the #1 fan of range contraction trading. Oswald also has a webinar on his site entitled "Principles of Trading Range Contraction". It's a good video but you have to register to view it.

You may also find Alan Farley's work on "coiled springs" and NR7s helpful. He goes in-depth on coiled springs in his book, The Master Swing Trader.

P.S. I almost forgot to answer the question of: How Narrow is Narrow?

For daytrading I like to find candles that are 1% or less than the price of the stock. So for a $20 stock any candles with a range of 20 cents or less really get my attention. In cases like those I just need small moves in the stock to rack up some nice gains. I'll typically pass on candles with a range much greater than 1% because they'll require a much larger move in the stock in order for me to get those large R-multiples.

For swing trading (looking at daily charts) I don't have a concrete answer on how narrow is narrow. But you could probably come up with some value based on the stock's average true range (ATR) or some other method. For example, you could define narrow as less than half a stocks ATR.

NRx (narrowest range on the last X candles) is always a good way to go as well, and it's easy to scan for.

Michelle B submits: I remember reading about a French housewife who when asked by her husband every morning what they would be having for dinner would always give the same tireless reply: "It depends on what jumps into my market basket." Some traders prefer to trade from a prepared watchlist. For me, trading is interesting because I do not know what will jump into my market basket. My preparation is my confidence in my ability to execute perceived opportunities according to my risk parameters.

Using the premarket top gainers/losers scan, I search the NASDAQ, NYSE, and AMEX market stalls for any tasty morsels and look for any stocks moving on the highs/lows scan. I also check Briefing.com and MarketWatch Newsfinder for news that has happened since the close of the previous trading day---identifying which stocks the market considers to be newsworthy is one of the many trading skills which can be developed.

In addition, I note via Briefing.com calendars such events as economic reports, earnings, Fedspeak, splits, upgrades/downgrades, and conferences. Usually, I have around ten to twelve trading candidates. Sometimes, there are none or a very small number. In that case, I patiently wait until the market has a new batch of fresh candidates---I will not make do with stale merchandise.

I list the stock symbols of my candidates and the important economic events for that day in my journal. Next to each stock symbol, I put the results of my basic Yahoo financial research---reason for price movement, float size, and short interest. I identify support and resistance on various timeframes. Sometimes, I check Yahoo Finance message boards and blogs to find out why a stock may be moving. Though in that case, I am always on my guard to weed the chaff from the wheat.

Then, I place my candidates on the chart pages of my Advanced Trading Platform (ATP). Four thirty-minute charts go on pages two through four. The main page contains the following windows: order entry, order status, real time portfolio, high/low scan, top gainers/losers scan, watchlist containing, in addition to my trading candidates, the major indices and technical indicators like VIX and TICK. Last but not least, is a chart window for the most promising candidate.

Discovering that ININ---set to gap up because it has given higher earnings' guidance---has a very small float and moderately high short interest, coupled with my knowledge that such stocks often move very strongly on low volume days, like Fridays, pre-holidays, and summer days, makes ININ the star attraction at this stage.

The market opens, and since I am particularly interested in seeing if ININ can trigger a trade, it gets top attention with a one-minute chart (above) on my main trading page. Meanwhile, I am flipping through my chart pages to see if anything is close to triggering---nothing yet. I check the one-minute chart of ININ, and I see the choppy pennant being formed at 10 A.M. The pennant range is around .25. That's enough of a sign for me to buy half of my lot as close to the bottom of this formation at 13.65. Buying a partial lot before the actual breakout happens makes sense in high-probability trades, especially if the desired lot size is substantial. I am set to buy the rest via an automatic buy stop if it can clear the top of the pennant at 13.87 on the 5 minute (see below). Trades moving strongly in the morning require monitoring via shorter time frames.

Fifteen minutes later, it does, and my full lot is in, average cost around 13.78. Just before the breakout, the pennant was around seventy-five percent completed and the volume dried up. Triangles/pennants need to be not completely formed or else they usually just fizzle out and do not support strong breakouts. The pause in volume is what my trading buddy calls a lullipop, a lull before the breakout, where the equilibrium between buyers and sellers end and the buyers take over.

Part of my prepurchase research of ININ was its support and resistance based on weekly price level resistance. Since the open of the July 24, 2006 weekly, bearish, high-volume candle was 14.81 (see below), therefore trapping lots of traders at that level, I determined the resistance to be around 14.80. The target of 14.80 also matches the length of the pole on which the flag formed on Friday's intraday charts, forming a measured move and confirming the resistance.

As the ININ trade continues, I am monitoring the price action on different timeframes, and have my mental stop under the flag for the first half lot at 13.53 and then just above the flag at 13.89 for the complete lot. Price is smoothly rising---each candle opens near but not with much overlapping to the close of the previous candle. I put my sell limit in for the full lot just under the target price. My offer gets lifted fairly easily, and I am out with a $1.00 move gotten in about 40 minutes. Since the risk was .25 for the complete lot, it was a +4R trade. For quickly moving stocks, I will often forgo hard stops, but it is a risk because my net connection could be interrupted. Then my carefully constructed trade will exist only in cyberspace floating without my pilotage. So, I am working on the artful science of consistently setting hard regular/trailing stops with upper triggers via bracketed orders to take care of that possibility.

My market basket is now empty until the next fresh thing jumps into it.

There seems to be a lot of controversial over the concept of R-Multiples. I've been seeing people complain about them for months now and I've been meaning to write a post about "R". I really wanted to do it last week but I'm glad I didn't get to it because this week a raging debate about R has popped up. Glenn, at DehTrader can serve as the poster child for the anti-R crew. Here's part of his recent rant against R-Multiples (emphasis is mine):

I post real numbers as opposed to R values, I always have. I like real numbers, I understand real numbers and I see truth in real numbers and I think the reader does too. As a reader of many blogs I find zero value in any post or summaries containing R values, I don't see any point in sharing that information. I suppose if I posted in R values I could look like a pretty good trader, but we all know I am a struggling trader. R can mean anything so why even bother with it... R stands for bullsh!t imo and that's my rant (that and ads haha). The best blogs out there post real numbers, Boogtser, JC (NYSE), the Kirkster all come to mind.

He's joined by folks like Paul who left this comment over on Ugly's post about R multiples:

I believe dollar values are more important than R value. I agree that the actual $ value is meaningless. However, R values are subjective and don’t give you a true idea on how successful the trade was. If you defined your risk at 15 cents and made 30 cents on the trade, while another person made 50 cents but decided his risk would be 50 cents, R values would say the guy who made 30 cents was more successful. I have a problem with that. It could very well be that the guy who only risked 15 cents is playing it too safe and his 2R gain was a bad trade.

So that gives you an idea of the anti-R sentiment. I'm going to explain why I think R-Multiples are so useful and why I use them in my trading and on this site.

What is R?

R is simply the dollar risk per trade. It's nothing but a reward-to-risk ratio. I first heard it called "R" in Van Tharp's book "Trade Your Way to Financial Freedom". In another of his books, "Financial Freedom Through Electronic Day Trading", Dr. Tharp reveals the great secret of trading:

The golden rule of trading is to keep losses at a level of 1 R as often as possible and to make profits that are high-R multiples.

You often hear (read) that traders should only look for trades with a reward/risk ratio of at least 2 or 3 to 1. Expressing your results in terms of how many times your risk allows you to easily see how well your trades measure up to such a standard. So when I look at my results in terms of multiples of R I can easily tell how good or bad the trades were. I like to think of R-Multiples as telling you the efficiency of your system.

So why not just use dollars?

Expressing my results in dollars would achieve the same result if I always risked the same amount of money. But what if I triple my account and therefore trade larger positions compared to when I started trading? Or what if I hit a rough spot and decide to cut my share size down while I ride out the storm? Then the dollar results won't easily tell me how trades from one period of time compared to another period of time. But if I use R making such comparisons is simple. Either my trades passed the risk / reward ratio test or they didn't. The actual number of dollars at risk doesn't matter, how many multiples of the dollars at risk does.

Along the same lines, recording trades in terms of R-Multiples allows you to easily calculate your system's expectancy. (Follow the link for why you should care about expectancy.)

Also, as Rx said:

talking and thinking in terms of R-multiple when you discuss about profits is an excellent approach - that by itself makes you focus on risk and money management - the actual "grail" to successful trading.

That is a very important point. Whenever I see people posting dollar returns, especially losses, that are all over the place the first thing I ask myself is "I wonder what his risk per trade is". It's almost a certainty that those traders aren't focusing on risk and as a result keep having huge losses. The mere fact that you have to define R and then place a stop to keep your loss to 1R is probably too constraining for those gamblers traders. Dr. Tharp says about determining your initial stop-loss point as soon as you enter a trade, which, by definition woud give you a 1 R loss:

This principle is so important that if you cannot follow it, then you might as well give up the idea of electronic day trading right away.

The reason I use R on the blog is because I don't want to discuss dollars or my account size on the site (as Ugly stated). That's nobody's business but mine. Also, it makes it easy for people to figure out what they could have made or lost on a trade with their own account size and risk per trade amount. If you see a trade that returned 3R all one has to do is plug in their dollar risk per trade to figure out what they could have made / lost.

To the R Haters

Let me address the "alleged" issues which I quoted above...

Glenn thinks that R is just some made up number and could mean anything. He likes "real" numbers. While it may be fun to see that somebody made $10,000 on a trade that in and of itself doesn't tell you how good that trade was. What if that person risked $30,000 to make that $10,000? Or what if they risked $1,000 to make that $10,000? Those are two very different trades. Sure they both made the same amount of money but isn't the second trade a much more efficient use of capital?

What if somebody is trading $500,000 lots to make $1,000 in profit? It may be nice to see somebody saying that they made $1,000 here and $1,000 there but damn(!) that's an inefficient use of capital. So while R could mean anything in terms of dollars, in my humble opinion what really matters is how many multiples of R were made or lost. That tells you the quality of a trade or system.

Glenn also states that if he reported his trades in terms of R he could appear to be a good trader. I'm sorry to tell him that's simply not the case. If you lost money that means your expectancy, which is just your average return expressed in R-Multiples, was negative.

Paul said that "R values are subjective and don’t give you a true idea on how successful the trade was". That is exactly wrong. R-multiples are the very thing that tells you exactly how successful a given trade was, if you choose to grade on a risk/reward basis.

Percentages vs. Dollars

This debate about R reminds me of a conversation I had a couple of weeks ago. I was in a presentation for Trade-Ideas' new tool, the Odds Maker. They were showing how you could backtest all these different scenarios with the tool. The results were expressed in average dollars won or lost. Another viewer and I asked about seeing the results in percentages. They kept saying that perhaps they would do that in a later revision. I kept harping on it because to me seeing the results in dollars was of little use for the way I size my positions

The argument from the presenter was that all you had to do was multiply the average dollar return by your average lot size to figure out how much money you could have made with a given system you were testing. I had to disagree because my lot size can vary drastically depending on how far away my stop loss is. Here's a situation which could be problematic -- I trade Google with a 2 point stop (which is only about half of a percent) and get lucky and make 6 points of profit. All of my other trades are on stocks under $50 with stops less than 50 cents. I could have some combination of winners and losers mixed in there... most of them probably well less than $6. That $6 gain may skew the results when presented as average dollars won. That's an over-simplification and there are all kinds of possible permutations. But I hope my point is clear that looking at the results in terms of average dollars won/lost may not tell accurately tell you the story.

So how can we make the results clearer? Simple, express them in percentages. That way, regardless of how many shares were traded or the prices of the stocks traded the results can be equalized across all the trades. I feel much better being able to say , "OK, this system would have returned X%" instead of "X number of points.

We debated the merits of each way of reporting for a few minutes and at one point somebody said, well , for this release we're aiming for the "lowest common denominator". In other words, the average person can't think in percentages, so we're just gonna report in points. I was like, F the average person, make it work the "right" way! The funny thing is that after debating all of that the software actually could express the results in percentage terms. We just had to switch a setting.

So my point of that little story is that I always prefer to think in terms of percentages in stead of points. I always see people talking about number of shares of point moves. For example, you might hear somebody exclaim "Google is up 5 points!!!" I don't see that as anything to get excited about. That just over a 1% move -- a normal fluctuation. You'll hear similar things from reporters talking excitedly about the Dow being up some triple-digit amount. The Nasdaq may actually be up a lot more on a percentage basis but they'll just say, eh, the Nasdaq is "just" up 30 points.

Looking at the percentages makes those kind of comparisons easier. R-Multiples do the same thing for traders. They can accurately compare their own trades and they can take another trader's results expressed in R and easily relate them to their own system.

Cameron just left a question for me in the comments to last night's recap. He asked:

I'm looking at your way of thinking about the market. One of the biggest questions I have, though, is regarding your stops.

You said you set your stop at just .20 over your short dollar amount. With all of the intra-day fluctuations, how can you know that it wouldn't have gone up, say, .25 and then nose-dived? It's a perfectly plausible situation.

Said another way, I can place a buy order with a stop, but we easily have 1% swings in a day. If a swing occurs, I have automatically sold for a lower price, only to see the market go back up. Therefore, I'm encouraged to set much broader stops. You set yours for only .20 over the buy price, though, which is why I'm confused.

I often get asked variations of this question, so I thought I'd put my answer in its own post. This was my response:

I couldn't have known that the stock wouldn't have shaken me out. But I also wouldn't have known if I'd have been shaken out if my stop was 50 cents or even a dollar higher. That narrow range candle just provided me with a tight stop. I like tight stops b/c they give me a better opportunity to get high R-multiple winners.

But to your point, how tight of a stop is too tight and will practically guarantee being stopped out? It's hard to answer that question but I generally shy away from any stop that's less than 0.5% of the stock price. I normally like my stops to be about 1% of the stock price. In general, I'm more comfortable with somewhat tighter stops on the QQQQ b/c of how liquid it is and b/c it's less volatile than an "in play" stock would be.

Also, note that I didn't pick a 20 cent stop out of thin air. I went with what the previous candle told me was resistance.

The key thing to remember is that I'm on the hunt for high R-Multiple winners. Because of the way I size my positions (the percent risk model), the tighter my stops the higher the potential reward. In other words if I think a stock has potential to move $2 I'd rather have a 20 cent stop than a $1 stop. Risking 20 cents to make 200 cents ($2) would give me a 10R winner vs only a 2R winner if I had risked $1. (If R was 1% of my equity that would be a 10% gain vs a 2% gain.) For a real example of this see the DGX trades I posted on my Narrow Range Bars post.

So there' a trade-off at work here between stop size and risk/reward multiples. The tighter the stop, the better chance you have of getting a high R winner but you also have a better chance of getting stopped out. Your win ratio is likely to be inversely proportional to how tight your stops are -- looser stops should lead to a higher win rate but also fewer high R-multiple winners.

It's also key that you don't pick your stops willy-nilly. I always use the support/resistance levels that the charts dictate to me.

There's some good advice in Alan Farley's latest article to help "home gamers" (Cramer's phrase) survive the current extra-volatile market. The whole article is worth reading but here are the four bits of advice:

I have four pieces of cautionary advice for my at-home brethren. Listen up, because it's no longer a question of whether you want to take real money out of the market, or just add a few bucks to the weekly shopping budget. These remedial steps must be taken if you want to survive long enough to take advantage of the real opportunity.

1. Wait for the Market Volatility Index (VIX) to drop below 40 and stay there for a week. Massive price swings require equally massive stop losses, which rarely justify the intended positions. Your only alternative is to stand aside and do nothing, no matter how much it hurts to watch others playing those big rallies and selloffs.
2. Forget overnight positions until the index futures stop gapping 2% or more every morning. These price jolts are great news when you're on the right side of the trade, but total devastation if you're on the wrong side. And guess what? You're not smart enough to predict overnight direction from day to day. Neither am I.
3. End your love affair with popular stocks that made you money during the last bull market. In November 2008, these are the issues that will trigger the most painful and unexpected reversals, which happen right after you're absolutely convinced your position is the right play. The bottom line: They see you coming, sucker.
4. Get control of the time element in your market strategy. You're getting killed because you have no patience and forgot how to sit on your hands when your trading edge isn't in play. Realistically, it could be months before the market works for you again. Would you rather wait it out and survive, or stay busy and get crushed?

My thoughts:

1. I like that part about watching the VIX. A few weeks ago I decided to cut my risk per trade (R) in half (from 1% to 0.5% of my equity) until the VIX breaks that top it's trying to put in. That means a close below 44. (Dr. Brett recently wrote about when traders should cut their risk.)
2. Tough to argue with what he said about overnight holds. This is certainly a market for intraday trades.
3. Reminds me of what I said about AAPL and GOOG back in July. Add DRYS and its ilk to the list too. It's better to just have an open mind and grab on to whichever stocks become the new leaders.
4. Yep, preservation of capital must be the top priority. If you can't sit out at least lower your risk.

I was recently asked about how I determine my "universe" of stocks. That is the stocks that I consider fair game to be traded. I use different filters for swing trading and day trading but they do have similar criteria. In both cases I'm trying to find liquid, volatile stocks. Once the filter's applied I then look for certain patterns / setups to trade.

For swing trade candidates I use the following criteria in TeleChart:

• Price Per Share: about $5 to Max
• 5 Day Average Volume: 75th percentile & higher (right now that works out to about 466,000 shares per day)
• Price Volatility: 25th percentile & higher
• Volume (Dollars) 5-Day: 75th percentile & higher

Here's a shot of the filter. The count column shows how many stocks are left after applying each row in the filter.

That typically narrows the market down to about 1,000 stocks. Then I look for setups among that set of stocks. I gave a lot of details about the setups I scan for back in the day.

For day trading I'm trying to find liquid stocks which are abnormally active and volatile. To do that I use the following two scans in Trade-Ideas:

Bearish Scanner's Filter:

• Min Price: $10
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: -0.5% (negative means down at least 0.5%)

Bullish Scanner's Filter:

• Min Price: $5
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: 1%

Sometimes I'll tweak these settings to give more or fewer alerts. I usually do that by adjusting the "Min Current Volume". Details about the actual alerts & setups I track for daytrading are on my "how I work" page. If you use Trade-Ideas you can paste the following link into the "Collaborate" feature and load that filter and all the alerts I track:

Bearish -- grab the URL of this link

Bullish -- grab the URL of this link

One thing I have always preached on the blog, less so on Twitter, where I bang out more trading ideas and market thoughts is that shorting stocks is hard. I think it’s harder than any aspect of learning the market. It’s dangerous. Let this morning be the only reminder you should EVER need.

Howard's right about shorting being difficult. I think it's so hard because it's not simply the opposite of going long. (I won't even go into the whole thing about your losses while short being theoretically infinite. Been there... done that. Use a stop to limit your losses, use proper position sizing, ONLY short LIQUID stocks and you'll be fine.) What makes shorting tricky is that bear moves often have violent (short-covering) rallies because the psychology of the crowd trading a down market is different than that of a bull market. You have to be quick on your feet when shorting. My motto is "stick & move".

Many traders love to buy breakouts in bull markets. (Whether that's actually a good strategy is debatable). In my experience swing trading, the opposite of that strategy, shorting breakdowns through resistance, will often lead you right into a snapback rally and, as MaoXian used to say "the quickest loss ever". That's why I often make note of all the people who are initiating shorts after the market has already fallen to a major support level. We saw that this week as I noted in some of the morning watchlists.

My contention is that if you were caught short Friday morning you should consider your losses as tuition paid to the school of hard knocks. Learn from that expensive lesson, take your losses and hopefully survive to trade another day. I'll stop short of saying that the losses were deserved but there were plenty of warnings to at least cover your shorts if not to get long. There are so many good sources of market information these days, both on the web and, yes, even on TV. I'm not saying to blindly follow somebody else's opinion but it can be helpful to see what others are doing based on what they see. Here are just a few of the recent warning signs:

• T2108 dropping below 20 -- Ah, good old reliable T2108. I've been watching it closely as we've sold off. On Wednesday I noted that it finally hit the point where wise shorts would want to cover. It pays to find a good overbought/oversold indicator and heed its warnings. (You may need different indicators or settings for different timeframes.) Sell at overbought and cover at oversold. A couple of years ago I decided to force myself to put some IRA money to work every time T2108 broke 20. It hasn't failed me yet.
• The Fed (Plunge Protection Team) has interfered announced stimulus packages around the July lows a couple times. On Thursday and earlier in the week there was talk of more PPT action.
• The PPT took action earlier in the week and last week. Just look at the orchestration of the LEH and AIG situations, etc.
• Extreme Volatility -- The moves all week were nothing short of violent. Positions were whipsawed all over the place. That in & of itself would be reason enough to lighten up on positions if not move to the sidelines. That kind of volatility is often a sign of a trend reversal, not of a continuation of the previous trend. That's why so many traders watch the VIX. Tons of people noted the spike in the VIX on Thursday.
• Corey from the 'Afraid to Trade' blog warned 'Use Extreme Caution in the Week Ahead.' He gave many good reasons, including the Federal Reserve interest rate decision, the quadruple witching options expiration and headline risk from troubled financial firms. His crystal ball was working well when he wrote "We could see a week ahead that will be discussed years later - as such, if you are a newer trader, it might be best to switch to simulation mode this week or use this week as a training experience, rather than risking real capital in an environment that could swing violently up and down due to market events scheduled to happen this week."
• Bullish Technical Divergences -- Dr. Brett pointed out some bullish divergences he was seeing in the market. Perhaps most important were what he called 'those fuzzy indicators' -- "Traffic on the blog is way up, reflecting trader uncertainty and desire for information. I just fielded my fourth media interview request in two days. During quiet and bullish market periods, I don't get four requests in a month." On Tuesday morning I also noted my blog's traffic spiked as did Barry Ritholtz. I've often joked about making some kind of sentiment indicator based on my site's traffic ebb & flow & referral logs. It's not a bad idea and I think it would be especially useful if it were based on a major financial site's traffic data.
• Dennis Gartman telling folks to "be small" -- Gartman was on CNBC's 'Fast Money' early in the week saying that he was scared of the market's movement and he was "being small" and planned to "get smaller". His advice to others was to "be small" in this market.
• Jeff Macke, also on 'Fast Money' was warning people not to play (trade) if they didn't understand the (changing) rules of the game. He was referring to all the headline risk from PPT action and the volatility caused by rumors -- many of which were spread by CNBC during the trading sessions. Ironically, Macke was kicking himself Thursday night for not following his own advice and getting caught short.

I fully believe that had the PPT not acted on Thursday night the market was *destined* to move higher in the short term on its own. Still being short on expiration Friday in this environment was just asking for trouble. So what's a trader to do? Like I said earlier, stick & move. I think it makes much more sense to short bounces back to a trendline or moving average. William O'Neil's book on short selling talks about using retracements to the 50 and/or 200-day moving average as a place to initiate shorts. By the time the stock (or whatever instrument you're trading) is extended from its moving average it's time to cover. Then you can wait for another bounce to reload.

The strategy to cover & reload makes more sense when you're short due to the fact that if you're dead right with your call on the stock dropping the most you can gain is 100%. The odds of that happening are slim and if it ever did happen you'd likely have to ride out some severe short squeezes. But you might be able to stick & move your way to more that a 100% gain by reloading multiple times. In theory, you could catch 15 10% moves lower in a stock that's falling, retracing & falling anew.

Short sellers need to be nimble, pick their entry and exit spots wisely and heed signs of impending reversals. Trying to short breakdowns of an already extended market is a sucker's play -- as is overstaying your welcome while short.

I got this question the other day:

When you short a stock at 30.00 and you want to place a stop limit in case it jumps up past your max loss of say, 31.00, how exactly is that order entered (everytime I try to do this, it sells (Ed: he means buys) my stock immediately - I'm doing something really wrong). Do I put a limit price of 31.00 and a stop price above that, i.e. 31.50 (thus saying it's ok to repurchase any shares between 31.00 and 31.50)?

This is one of the trickier orders to place. In that situation you would want to enter a buy stop limit order (buy on stop with a limit). The stop would be at 31.00 and the limit would be 31.50. Here's a screen shot of a similar order to cover QQQQ with a buy stop @ 47 with a limit of 47.10:

What would happen here is that once the stop price of 47.00 is hit a buy limit order of 47.10 would be entered. As with any limit order, you're guaranteed price but not execution, so you may not be filled. If you wanted to be sure that you got out (covered) just use a buy stop of $47.00.

In case you missed it in the links section the other day, be sure to read the interview with Michelle B over on the StockTickr blog.

The other day Downtown Trader asked me how I handled trailing my stop on that DGX short. I pretty much followed exactly what's in my post about how I take partial profits. I moved the stop to breakeven once my gain was equal to my initial risk of 50 cents per share. As it dropped further I moved the stop to lock in one-third of my profit. Why one-third? It's a value I came up with which I think allows a stock to have a normal retracement without stopping me out. It's sort of an arbitrary level, although I got the idea from the Fibonacci levels. I've found that method to work reasonably well after much trial and error with trailing stops. I read somewhere that you can consider a retracement to be normal as long as it didn't break the 62% Fib level. So I figured that I could trail my stops somewhere around that level to allow stocks to have their normal fluctuations. I generally get more aggressive with the trailing stops as the day wears on. I'll move them to 50% by 3:00 and just keep pushing them until the close -- assuming I don't bail before then like I did with DGX.

The other day I got the following email about narrow range bars:

I was going through your site and others and I keep on hearing, no entry on narrow range bars. What is a narrow range bar, and could you send me a picture of one?

A narrow range bar (candle) is simply a bar which has a range from high to low that's much less than the average bar for a given equity. You'll often see people, like Dr. Sen, tracking NR7 bars. NR7 means the narrowest range of the last seven bars. For example, here are some NR7s from last Friday:

So now that you know what narrow range candles are, why should you care about them? Since stocks generally cycle between periods of low volatility and high volatility it can be beneficial to identify when the high volatility times are approaching. Often times, especially for trending charts, narrow range periods -- periods of range contraction -- will presage range expansion. You can think of a stocks movement like a spring / slinky -- it need to recoil (contract) in order to build up enough potential energy for its next expansion. However, it's important thing to note that range contraction doesn't tell you the direction of the impending expansion only that an expansion may be coming. It's up to the trader to use other means (indicators, common sense?) to catch the expansion in the right direction.

One of the nice things about trading against narrow range bars is that they give you a tight stop. If you size your positions like I do, the smaller the stop means you can buy (or short) more shares. It's the trades with the really tight stops that let you make those huge R-Multiple trades that often make or break a trader. Here's an example from a trade I took last week.

I'm going to use the actual entries taken by myself and by Butterboy but instead of using my actual exits, I'll assume I held until the close. So let's say my equity is $100,000 and my R, the amount I want to risk per trade, is 1% or $1,000. Given that, here's the first trade, based off of my actual entry parameters:

• Entry at 53.96
• Stop was 50 cents higher so I could short 1000/.50 or 2,000 shares
• DGX closed at 49.64, which is $4.32 lower. So that trade returned $8,640 or 8.64 times my risk (8.64R)

Here's the DGX trade using Butterboy's entry criteria. He used a 15 minute chart and entered lower than I did but his stop was also much tighter:

• Entry at 53.40
• I believe his stop was 21 cents higher, just above the 11:30 bar which he entered against. Given that one could short 1,000/.21 or just over 4,700 shares.
• DGX closed at $3.76 beneath the entry so that trade returned $17,672 or 17.67 times the initial risk (17.64R).

It should be clear why so many of us like narrow range bars and why I scan for them. That second trade had a worse (lower) entry but because of its tighter stop it was almost twice as profitable. It also shows why I love the percent risk position sizing model.

Here's an article on range contraction written by Oswald S. Castillo, TheStockStalker and very possibly the #1 fan of range contraction trading. Oswald also has a webinar on his site entitled "Principles of Trading Range Contraction". It's a good video but you have to register to view it.

You may also find Alan Farley's work on "coiled springs" and NR7s helpful. He goes in-depth on coiled springs in his book, The Master Swing Trader.

P.S. I almost forgot to answer the question of: How Narrow is Narrow?

For daytrading I like to find candles that are 1% or less than the price of the stock. So for a $20 stock any candles with a range of 20 cents or less really get my attention. In cases like those I just need small moves in the stock to rack up some nice gains. I'll typically pass on candles with a range much greater than 1% because they'll require a much larger move in the stock in order for me to get those large R-multiples.

For swing trading (looking at daily charts) I don't have a concrete answer on how narrow is narrow. But you could probably come up with some value based on the stock's average true range (ATR) or some other method. For example, you could define narrow as less than half a stocks ATR.

NRx (narrowest range on the last X candles) is always a good way to go as well, and it's easy to scan for.

Michelle B submits: I remember reading about a French housewife who when asked by her husband every morning what they would be having for dinner would always give the same tireless reply: "It depends on what jumps into my market basket." Some traders prefer to trade from a prepared watchlist. For me, trading is interesting because I do not know what will jump into my market basket. My preparation is my confidence in my ability to execute perceived opportunities according to my risk parameters.

Using the premarket top gainers/losers scan, I search the NASDAQ, NYSE, and AMEX market stalls for any tasty morsels and look for any stocks moving on the highs/lows scan. I also check Briefing.com and MarketWatch Newsfinder for news that has happened since the close of the previous trading day---identifying which stocks the market considers to be newsworthy is one of the many trading skills which can be developed.

In addition, I note via Briefing.com calendars such events as economic reports, earnings, Fedspeak, splits, upgrades/downgrades, and conferences. Usually, I have around ten to twelve trading candidates. Sometimes, there are none or a very small number. In that case, I patiently wait until the market has a new batch of fresh candidates---I will not make do with stale merchandise.

I list the stock symbols of my candidates and the important economic events for that day in my journal. Next to each stock symbol, I put the results of my basic Yahoo financial research---reason for price movement, float size, and short interest. I identify support and resistance on various timeframes. Sometimes, I check Yahoo Finance message boards and blogs to find out why a stock may be moving. Though in that case, I am always on my guard to weed the chaff from the wheat.

Then, I place my candidates on the chart pages of my Advanced Trading Platform (ATP). Four thirty-minute charts go on pages two through four. The main page contains the following windows: order entry, order status, real time portfolio, high/low scan, top gainers/losers scan, watchlist containing, in addition to my trading candidates, the major indices and technical indicators like VIX and TICK. Last but not least, is a chart window for the most promising candidate.

Discovering that ININ---set to gap up because it has given higher earnings' guidance---has a very small float and moderately high short interest, coupled with my knowledge that such stocks often move very strongly on low volume days, like Fridays, pre-holidays, and summer days, makes ININ the star attraction at this stage.

The market opens, and since I am particularly interested in seeing if ININ can trigger a trade, it gets top attention with a one-minute chart (above) on my main trading page. Meanwhile, I am flipping through my chart pages to see if anything is close to triggering---nothing yet. I check the one-minute chart of ININ, and I see the choppy pennant being formed at 10 A.M. The pennant range is around .25. That's enough of a sign for me to buy half of my lot as close to the bottom of this formation at 13.65. Buying a partial lot before the actual breakout happens makes sense in high-probability trades, especially if the desired lot size is substantial. I am set to buy the rest via an automatic buy stop if it can clear the top of the pennant at 13.87 on the 5 minute (see below). Trades moving strongly in the morning require monitoring via shorter time frames.

Fifteen minutes later, it does, and my full lot is in, average cost around 13.78. Just before the breakout, the pennant was around seventy-five percent completed and the volume dried up. Triangles/pennants need to be not completely formed or else they usually just fizzle out and do not support strong breakouts. The pause in volume is what my trading buddy calls a lullipop, a lull before the breakout, where the equilibrium between buyers and sellers end and the buyers take over.

Part of my prepurchase research of ININ was its support and resistance based on weekly price level resistance. Since the open of the July 24, 2006 weekly, bearish, high-volume candle was 14.81 (see below), therefore trapping lots of traders at that level, I determined the resistance to be around 14.80. The target of 14.80 also matches the length of the pole on which the flag formed on Friday's intraday charts, forming a measured move and confirming the resistance.

As the ININ trade continues, I am monitoring the price action on different timeframes, and have my mental stop under the flag for the first half lot at 13.53 and then just above the flag at 13.89 for the complete lot. Price is smoothly rising---each candle opens near but not with much overlapping to the close of the previous candle. I put my sell limit in for the full lot just under the target price. My offer gets lifted fairly easily, and I am out with a $1.00 move gotten in about 40 minutes. Since the risk was .25 for the complete lot, it was a +4R trade. For quickly moving stocks, I will often forgo hard stops, but it is a risk because my net connection could be interrupted. Then my carefully constructed trade will exist only in cyberspace floating without my pilotage. So, I am working on the artful science of consistently setting hard regular/trailing stops with upper triggers via bracketed orders to take care of that possibility.

My market basket is now empty until the next fresh thing jumps into it.

There seems to be a lot of controversial over the concept of R-Multiples. I've been seeing people complain about them for months now and I've been meaning to write a post about "R". I really wanted to do it last week but I'm glad I didn't get to it because this week a raging debate about R has popped up. Glenn, at DehTrader can serve as the poster child for the anti-R crew. Here's part of his recent rant against R-Multiples (emphasis is mine):

I post real numbers as opposed to R values, I always have. I like real numbers, I understand real numbers and I see truth in real numbers and I think the reader does too. As a reader of many blogs I find zero value in any post or summaries containing R values, I don't see any point in sharing that information. I suppose if I posted in R values I could look like a pretty good trader, but we all know I am a struggling trader. R can mean anything so why even bother with it... R stands for bullsh!t imo and that's my rant (that and ads haha). The best blogs out there post real numbers, Boogtser, JC (NYSE), the Kirkster all come to mind.

He's joined by folks like Paul who left this comment over on Ugly's post about R multiples:

I believe dollar values are more important than R value. I agree that the actual $ value is meaningless. However, R values are subjective and don’t give you a true idea on how successful the trade was. If you defined your risk at 15 cents and made 30 cents on the trade, while another person made 50 cents but decided his risk would be 50 cents, R values would say the guy who made 30 cents was more successful. I have a problem with that. It could very well be that the guy who only risked 15 cents is playing it too safe and his 2R gain was a bad trade.

So that gives you an idea of the anti-R sentiment. I'm going to explain why I think R-Multiples are so useful and why I use them in my trading and on this site.

What is R?

R is simply the dollar risk per trade. It's nothing but a reward-to-risk ratio. I first heard it called "R" in Van Tharp's book "Trade Your Way to Financial Freedom". In another of his books, "Financial Freedom Through Electronic Day Trading", Dr. Tharp reveals the great secret of trading:

The golden rule of trading is to keep losses at a level of 1 R as often as possible and to make profits that are high-R multiples.

You often hear (read) that traders should only look for trades with a reward/risk ratio of at least 2 or 3 to 1. Expressing your results in terms of how many times your risk allows you to easily see how well your trades measure up to such a standard. So when I look at my results in terms of multiples of R I can easily tell how good or bad the trades were. I like to think of R-Multiples as telling you the efficiency of your system.

So why not just use dollars?

Expressing my results in dollars would achieve the same result if I always risked the same amount of money. But what if I triple my account and therefore trade larger positions compared to when I started trading? Or what if I hit a rough spot and decide to cut my share size down while I ride out the storm? Then the dollar results won't easily tell me how trades from one period of time compared to another period of time. But if I use R making such comparisons is simple. Either my trades passed the risk / reward ratio test or they didn't. The actual number of dollars at risk doesn't matter, how many multiples of the dollars at risk does.

Along the same lines, recording trades in terms of R-Multiples allows you to easily calculate your system's expectancy. (Follow the link for why you should care about expectancy.)

Also, as Rx said:

talking and thinking in terms of R-multiple when you discuss about profits is an excellent approach - that by itself makes you focus on risk and money management - the actual "grail" to successful trading.

That is a very important point. Whenever I see people posting dollar returns, especially losses, that are all over the place the first thing I ask myself is "I wonder what his risk per trade is". It's almost a certainty that those traders aren't focusing on risk and as a result keep having huge losses. The mere fact that you have to define R and then place a stop to keep your loss to 1R is probably too constraining for those gamblers traders. Dr. Tharp says about determining your initial stop-loss point as soon as you enter a trade, which, by definition woud give you a 1 R loss:

This principle is so important that if you cannot follow it, then you might as well give up the idea of electronic day trading right away.

The reason I use R on the blog is because I don't want to discuss dollars or my account size on the site (as Ugly stated). That's nobody's business but mine. Also, it makes it easy for people to figure out what they could have made or lost on a trade with their own account size and risk per trade amount. If you see a trade that returned 3R all one has to do is plug in their dollar risk per trade to figure out what they could have made / lost.

To the R Haters

Let me address the "alleged" issues which I quoted above...

Glenn thinks that R is just some made up number and could mean anything. He likes "real" numbers. While it may be fun to see that somebody made $10,000 on a trade that in and of itself doesn't tell you how good that trade was. What if that person risked $30,000 to make that $10,000? Or what if they risked $1,000 to make that $10,000? Those are two very different trades. Sure they both made the same amount of money but isn't the second trade a much more efficient use of capital?

What if somebody is trading $500,000 lots to make $1,000 in profit? It may be nice to see somebody saying that they made $1,000 here and $1,000 there but damn(!) that's an inefficient use of capital. So while R could mean anything in terms of dollars, in my humble opinion what really matters is how many multiples of R were made or lost. That tells you the quality of a trade or system.

Glenn also states that if he reported his trades in terms of R he could appear to be a good trader. I'm sorry to tell him that's simply not the case. If you lost money that means your expectancy, which is just your average return expressed in R-Multiples, was negative.

Paul said that "R values are subjective and don’t give you a true idea on how successful the trade was". That is exactly wrong. R-multiples are the very thing that tells you exactly how successful a given trade was, if you choose to grade on a risk/reward basis.

Percentages vs. Dollars

This debate about R reminds me of a conversation I had a couple of weeks ago. I was in a presentation for Trade-Ideas' new tool, the Odds Maker. They were showing how you could backtest all these different scenarios with the tool. The results were expressed in average dollars won or lost. Another viewer and I asked about seeing the results in percentages. They kept saying that perhaps they would do that in a later revision. I kept harping on it because to me seeing the results in dollars was of little use for the way I size my positions

The argument from the presenter was that all you had to do was multiply the average dollar return by your average lot size to figure out how much money you could have made with a given system you were testing. I had to disagree because my lot size can vary drastically depending on how far away my stop loss is. Here's a situation which could be problematic -- I trade Google with a 2 point stop (which is only about half of a percent) and get lucky and make 6 points of profit. All of my other trades are on stocks under $50 with stops less than 50 cents. I could have some combination of winners and losers mixed in there... most of them probably well less than $6. That $6 gain may skew the results when presented as average dollars won. That's an over-simplification and there are all kinds of possible permutations. But I hope my point is clear that looking at the results in terms of average dollars won/lost may not tell accurately tell you the story.

So how can we make the results clearer? Simple, express them in percentages. That way, regardless of how many shares were traded or the prices of the stocks traded the results can be equalized across all the trades. I feel much better being able to say , "OK, this system would have returned X%" instead of "X number of points.

We debated the merits of each way of reporting for a few minutes and at one point somebody said, well , for this release we're aiming for the "lowest common denominator". In other words, the average person can't think in percentages, so we're just gonna report in points. I was like, F the average person, make it work the "right" way! The funny thing is that after debating all of that the software actually could express the results in percentage terms. We just had to switch a setting.

So my point of that little story is that I always prefer to think in terms of percentages in stead of points. I always see people talking about number of shares of point moves. For example, you might hear somebody exclaim "Google is up 5 points!!!" I don't see that as anything to get excited about. That just over a 1% move -- a normal fluctuation. You'll hear similar things from reporters talking excitedly about the Dow being up some triple-digit amount. The Nasdaq may actually be up a lot more on a percentage basis but they'll just say, eh, the Nasdaq is "just" up 30 points.

Looking at the percentages makes those kind of comparisons easier. R-Multiples do the same thing for traders. They can accurately compare their own trades and they can take another trader's results expressed in R and easily relate them to their own system.

Cameron just left a question for me in the comments to last night's recap. He asked:

I'm looking at your way of thinking about the market. One of the biggest questions I have, though, is regarding your stops.

You said you set your stop at just .20 over your short dollar amount. With all of the intra-day fluctuations, how can you know that it wouldn't have gone up, say, .25 and then nose-dived? It's a perfectly plausible situation.

Said another way, I can place a buy order with a stop, but we easily have 1% swings in a day. If a swing occurs, I have automatically sold for a lower price, only to see the market go back up. Therefore, I'm encouraged to set much broader stops. You set yours for only .20 over the buy price, though, which is why I'm confused.

I often get asked variations of this question, so I thought I'd put my answer in its own post. This was my response:

I couldn't have known that the stock wouldn't have shaken me out. But I also wouldn't have known if I'd have been shaken out if my stop was 50 cents or even a dollar higher. That narrow range candle just provided me with a tight stop. I like tight stops b/c they give me a better opportunity to get high R-multiple winners.

But to your point, how tight of a stop is too tight and will practically guarantee being stopped out? It's hard to answer that question but I generally shy away from any stop that's less than 0.5% of the stock price. I normally like my stops to be about 1% of the stock price. In general, I'm more comfortable with somewhat tighter stops on the QQQQ b/c of how liquid it is and b/c it's less volatile than an "in play" stock would be.

Also, note that I didn't pick a 20 cent stop out of thin air. I went with what the previous candle told me was resistance.

The key thing to remember is that I'm on the hunt for high R-Multiple winners. Because of the way I size my positions (the percent risk model), the tighter my stops the higher the potential reward. In other words if I think a stock has potential to move $2 I'd rather have a 20 cent stop than a $1 stop. Risking 20 cents to make 200 cents ($2) would give me a 10R winner vs only a 2R winner if I had risked $1. (If R was 1% of my equity that would be a 10% gain vs a 2% gain.) For a real example of this see the DGX trades I posted on my Narrow Range Bars post.

So there' a trade-off at work here between stop size and risk/reward multiples. The tighter the stop, the better chance you have of getting a high R winner but you also have a better chance of getting stopped out. Your win ratio is likely to be inversely proportional to how tight your stops are -- looser stops should lead to a higher win rate but also fewer high R-multiple winners.

It's also key that you don't pick your stops willy-nilly. I always use the support/resistance levels that the charts dictate to me.

There's some good advice in Alan Farley's latest article to help "home gamers" (Cramer's phrase) survive the current extra-volatile market. The whole article is worth reading but here are the four bits of advice:

I have four pieces of cautionary advice for my at-home brethren. Listen up, because it's no longer a question of whether you want to take real money out of the market, or just add a few bucks to the weekly shopping budget. These remedial steps must be taken if you want to survive long enough to take advantage of the real opportunity.

1. Wait for the Market Volatility Index (VIX) to drop below 40 and stay there for a week. Massive price swings require equally massive stop losses, which rarely justify the intended positions. Your only alternative is to stand aside and do nothing, no matter how much it hurts to watch others playing those big rallies and selloffs.
2. Forget overnight positions until the index futures stop gapping 2% or more every morning. These price jolts are great news when you're on the right side of the trade, but total devastation if you're on the wrong side. And guess what? You're not smart enough to predict overnight direction from day to day. Neither am I.
3. End your love affair with popular stocks that made you money during the last bull market. In November 2008, these are the issues that will trigger the most painful and unexpected reversals, which happen right after you're absolutely convinced your position is the right play. The bottom line: They see you coming, sucker.
4. Get control of the time element in your market strategy. You're getting killed because you have no patience and forgot how to sit on your hands when your trading edge isn't in play. Realistically, it could be months before the market works for you again. Would you rather wait it out and survive, or stay busy and get crushed?

My thoughts:

1. I like that part about watching the VIX. A few weeks ago I decided to cut my risk per trade (R) in half (from 1% to 0.5% of my equity) until the VIX breaks that top it's trying to put in. That means a close below 44. (Dr. Brett recently wrote about when traders should cut their risk.)
2. Tough to argue with what he said about overnight holds. This is certainly a market for intraday trades.
3. Reminds me of what I said about AAPL and GOOG back in July. Add DRYS and its ilk to the list too. It's better to just have an open mind and grab on to whichever stocks become the new leaders.
4. Yep, preservation of capital must be the top priority. If you can't sit out at least lower your risk.

I was recently asked about how I determine my "universe" of stocks. That is the stocks that I consider fair game to be traded. I use different filters for swing trading and day trading but they do have similar criteria. In both cases I'm trying to find liquid, volatile stocks. Once the filter's applied I then look for certain patterns / setups to trade.

For swing trade candidates I use the following criteria in TeleChart:

• Price Per Share: about $5 to Max
• 5 Day Average Volume: 75th percentile & higher (right now that works out to about 466,000 shares per day)
• Price Volatility: 25th percentile & higher
• Volume (Dollars) 5-Day: 75th percentile & higher

Here's a shot of the filter. The count column shows how many stocks are left after applying each row in the filter.

That typically narrows the market down to about 1,000 stocks. Then I look for setups among that set of stocks. I gave a lot of details about the setups I scan for back in the day.

For day trading I'm trying to find liquid stocks which are abnormally active and volatile. To do that I use the following two scans in Trade-Ideas:

Bearish Scanner's Filter:

• Min Price: $10
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: -0.5% (negative means down at least 0.5%)

Bullish Scanner's Filter:

• Min Price: $5
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: 1%

Sometimes I'll tweak these settings to give more or fewer alerts. I usually do that by adjusting the "Min Current Volume". Details about the actual alerts & setups I track for daytrading are on my "how I work" page. If you use Trade-Ideas you can paste the following link into the "Collaborate" feature and load that filter and all the alerts I track:

Bearish -- grab the URL of this link

Bullish -- grab the URL of this link

One thing I have always preached on the blog, less so on Twitter, where I bang out more trading ideas and market thoughts is that shorting stocks is hard. I think it’s harder than any aspect of learning the market. It’s dangerous. Let this morning be the only reminder you should EVER need.

Howard's right about shorting being difficult. I think it's so hard because it's not simply the opposite of going long. (I won't even go into the whole thing about your losses while short being theoretically infinite. Been there... done that. Use a stop to limit your losses, use proper position sizing, ONLY short LIQUID stocks and you'll be fine.) What makes shorting tricky is that bear moves often have violent (short-covering) rallies because the psychology of the crowd trading a down market is different than that of a bull market. You have to be quick on your feet when shorting. My motto is "stick & move".

Many traders love to buy breakouts in bull markets. (Whether that's actually a good strategy is debatable). In my experience swing trading, the opposite of that strategy, shorting breakdowns through resistance, will often lead you right into a snapback rally and, as MaoXian used to say "the quickest loss ever". That's why I often make note of all the people who are initiating shorts after the market has already fallen to a major support level. We saw that this week as I noted in some of the morning watchlists.

My contention is that if you were caught short Friday morning you should consider your losses as tuition paid to the school of hard knocks. Learn from that expensive lesson, take your losses and hopefully survive to trade another day. I'll stop short of saying that the losses were deserved but there were plenty of warnings to at least cover your shorts if not to get long. There are so many good sources of market information these days, both on the web and, yes, even on TV. I'm not saying to blindly follow somebody else's opinion but it can be helpful to see what others are doing based on what they see. Here are just a few of the recent warning signs:

• T2108 dropping below 20 -- Ah, good old reliable T2108. I've been watching it closely as we've sold off. On Wednesday I noted that it finally hit the point where wise shorts would want to cover. It pays to find a good overbought/oversold indicator and heed its warnings. (You may need different indicators or settings for different timeframes.) Sell at overbought and cover at oversold. A couple of years ago I decided to force myself to put some IRA money to work every time T2108 broke 20. It hasn't failed me yet.
• The Fed (Plunge Protection Team) has interfered announced stimulus packages around the July lows a couple times. On Thursday and earlier in the week there was talk of more PPT action.
• The PPT took action earlier in the week and last week. Just look at the orchestration of the LEH and AIG situations, etc.
• Extreme Volatility -- The moves all week were nothing short of violent. Positions were whipsawed all over the place. That in & of itself would be reason enough to lighten up on positions if not move to the sidelines. That kind of volatility is often a sign of a trend reversal, not of a continuation of the previous trend. That's why so many traders watch the VIX. Tons of people noted the spike in the VIX on Thursday.
• Corey from the 'Afraid to Trade' blog warned 'Use Extreme Caution in the Week Ahead.' He gave many good reasons, including the Federal Reserve interest rate decision, the quadruple witching options expiration and headline risk from troubled financial firms. His crystal ball was working well when he wrote "We could see a week ahead that will be discussed years later - as such, if you are a newer trader, it might be best to switch to simulation mode this week or use this week as a training experience, rather than risking real capital in an environment that could swing violently up and down due to market events scheduled to happen this week."
• Bullish Technical Divergences -- Dr. Brett pointed out some bullish divergences he was seeing in the market. Perhaps most important were what he called 'those fuzzy indicators' -- "Traffic on the blog is way up, reflecting trader uncertainty and desire for information. I just fielded my fourth media interview request in two days. During quiet and bullish market periods, I don't get four requests in a month." On Tuesday morning I also noted my blog's traffic spiked as did Barry Ritholtz. I've often joked about making some kind of sentiment indicator based on my site's traffic ebb & flow & referral logs. It's not a bad idea and I think it would be especially useful if it were based on a major financial site's traffic data.
• Dennis Gartman telling folks to "be small" -- Gartman was on CNBC's 'Fast Money' early in the week saying that he was scared of the market's movement and he was "being small" and planned to "get smaller". His advice to others was to "be small" in this market.
• Jeff Macke, also on 'Fast Money' was warning people not to play (trade) if they didn't understand the (changing) rules of the game. He was referring to all the headline risk from PPT action and the volatility caused by rumors -- many of which were spread by CNBC during the trading sessions. Ironically, Macke was kicking himself Thursday night for not following his own advice and getting caught short.

I fully believe that had the PPT not acted on Thursday night the market was *destined* to move higher in the short term on its own. Still being short on expiration Friday in this environment was just asking for trouble. So what's a trader to do? Like I said earlier, stick & move. I think it makes much more sense to short bounces back to a trendline or moving average. William O'Neil's book on short selling talks about using retracements to the 50 and/or 200-day moving average as a place to initiate shorts. By the time the stock (or whatever instrument you're trading) is extended from its moving average it's time to cover. Then you can wait for another bounce to reload.

The strategy to cover & reload makes more sense when you're short due to the fact that if you're dead right with your call on the stock dropping the most you can gain is 100%. The odds of that happening are slim and if it ever did happen you'd likely have to ride out some severe short squeezes. But you might be able to stick & move your way to more that a 100% gain by reloading multiple times. In theory, you could catch 15 10% moves lower in a stock that's falling, retracing & falling anew.

Short sellers need to be nimble, pick their entry and exit spots wisely and heed signs of impending reversals. Trying to short breakdowns of an already extended market is a sucker's play -- as is overstaying your welcome while short.

I got this question the other day:

When you short a stock at 30.00 and you want to place a stop limit in case it jumps up past your max loss of say, 31.00, how exactly is that order entered (everytime I try to do this, it sells (Ed: he means buys) my stock immediately - I'm doing something really wrong). Do I put a limit price of 31.00 and a stop price above that, i.e. 31.50 (thus saying it's ok to repurchase any shares between 31.00 and 31.50)?

This is one of the trickier orders to place. In that situation you would want to enter a buy stop limit order (buy on stop with a limit). The stop would be at 31.00 and the limit would be 31.50. Here's a screen shot of a similar order to cover QQQQ with a buy stop @ 47 with a limit of 47.10:

What would happen here is that once the stop price of 47.00 is hit a buy limit order of 47.10 would be entered. As with any limit order, you're guaranteed price but not execution, so you may not be filled. If you wanted to be sure that you got out (covered) just use a buy stop of $47.00.

In case you missed it in the links section the other day, be sure to read the interview with Michelle B over on the StockTickr blog.

The other day Downtown Trader asked me how I handled trailing my stop on that DGX short. I pretty much followed exactly what's in my post about how I take partial profits. I moved the stop to breakeven once my gain was equal to my initial risk of 50 cents per share. As it dropped further I moved the stop to lock in one-third of my profit. Why one-third? It's a value I came up with which I think allows a stock to have a normal retracement without stopping me out. It's sort of an arbitrary level, although I got the idea from the Fibonacci levels. I've found that method to work reasonably well after much trial and error with trailing stops. I read somewhere that you can consider a retracement to be normal as long as it didn't break the 62% Fib level. So I figured that I could trail my stops somewhere around that level to allow stocks to have their normal fluctuations. I generally get more aggressive with the trailing stops as the day wears on. I'll move them to 50% by 3:00 and just keep pushing them until the close -- assuming I don't bail before then like I did with DGX.

The other day I got the following email about narrow range bars:

I was going through your site and others and I keep on hearing, no entry on narrow range bars. What is a narrow range bar, and could you send me a picture of one?

A narrow range bar (candle) is simply a bar which has a range from high to low that's much less than the average bar for a given equity. You'll often see people, like Dr. Sen, tracking NR7 bars. NR7 means the narrowest range of the last seven bars. For example, here are some NR7s from last Friday:

So now that you know what narrow range candles are, why should you care about them? Since stocks generally cycle between periods of low volatility and high volatility it can be beneficial to identify when the high volatility times are approaching. Often times, especially for trending charts, narrow range periods -- periods of range contraction -- will presage range expansion. You can think of a stocks movement like a spring / slinky -- it need to recoil (contract) in order to build up enough potential energy for its next expansion. However, it's important thing to note that range contraction doesn't tell you the direction of the impending expansion only that an expansion may be coming. It's up to the trader to use other means (indicators, common sense?) to catch the expansion in the right direction.

One of the nice things about trading against narrow range bars is that they give you a tight stop. If you size your positions like I do, the smaller the stop means you can buy (or short) more shares. It's the trades with the really tight stops that let you make those huge R-Multiple trades that often make or break a trader. Here's an example from a trade I took last week.

I'm going to use the actual entries taken by myself and by Butterboy but instead of using my actual exits, I'll assume I held until the close. So let's say my equity is $100,000 and my R, the amount I want to risk per trade, is 1% or $1,000. Given that, here's the first trade, based off of my actual entry parameters:

• Entry at 53.96
• Stop was 50 cents higher so I could short 1000/.50 or 2,000 shares
• DGX closed at 49.64, which is $4.32 lower. So that trade returned $8,640 or 8.64 times my risk (8.64R)

Here's the DGX trade using Butterboy's entry criteria. He used a 15 minute chart and entered lower than I did but his stop was also much tighter:

• Entry at 53.40
• I believe his stop was 21 cents higher, just above the 11:30 bar which he entered against. Given that one could short 1,000/.21 or just over 4,700 shares.
• DGX closed at $3.76 beneath the entry so that trade returned $17,672 or 17.67 times the initial risk (17.64R).

It should be clear why so many of us like narrow range bars and why I scan for them. That second trade had a worse (lower) entry but because of its tighter stop it was almost twice as profitable. It also shows why I love the percent risk position sizing model.

Here's an article on range contraction written by Oswald S. Castillo, TheStockStalker and very possibly the #1 fan of range contraction trading. Oswald also has a webinar on his site entitled "Principles of Trading Range Contraction". It's a good video but you have to register to view it.

You may also find Alan Farley's work on "coiled springs" and NR7s helpful. He goes in-depth on coiled springs in his book, The Master Swing Trader.

P.S. I almost forgot to answer the question of: How Narrow is Narrow?

For daytrading I like to find candles that are 1% or less than the price of the stock. So for a $20 stock any candles with a range of 20 cents or less really get my attention. In cases like those I just need small moves in the stock to rack up some nice gains. I'll typically pass on candles with a range much greater than 1% because they'll require a much larger move in the stock in order for me to get those large R-multiples.

For swing trading (looking at daily charts) I don't have a concrete answer on how narrow is narrow. But you could probably come up with some value based on the stock's average true range (ATR) or some other method. For example, you could define narrow as less than half a stocks ATR.

NRx (narrowest range on the last X candles) is always a good way to go as well, and it's easy to scan for.

Michelle B submits: I remember reading about a French housewife who when asked by her husband every morning what they would be having for dinner would always give the same tireless reply: "It depends on what jumps into my market basket." Some traders prefer to trade from a prepared watchlist. For me, trading is interesting because I do not know what will jump into my market basket. My preparation is my confidence in my ability to execute perceived opportunities according to my risk parameters.

Using the premarket top gainers/losers scan, I search the NASDAQ, NYSE, and AMEX market stalls for any tasty morsels and look for any stocks moving on the highs/lows scan. I also check Briefing.com and MarketWatch Newsfinder for news that has happened since the close of the previous trading day---identifying which stocks the market considers to be newsworthy is one of the many trading skills which can be developed.

In addition, I note via Briefing.com calendars such events as economic reports, earnings, Fedspeak, splits, upgrades/downgrades, and conferences. Usually, I have around ten to twelve trading candidates. Sometimes, there are none or a very small number. In that case, I patiently wait until the market has a new batch of fresh candidates---I will not make do with stale merchandise.

I list the stock symbols of my candidates and the important economic events for that day in my journal. Next to each stock symbol, I put the results of my basic Yahoo financial research---reason for price movement, float size, and short interest. I identify support and resistance on various timeframes. Sometimes, I check Yahoo Finance message boards and blogs to find out why a stock may be moving. Though in that case, I am always on my guard to weed the chaff from the wheat.

Then, I place my candidates on the chart pages of my Advanced Trading Platform (ATP). Four thirty-minute charts go on pages two through four. The main page contains the following windows: order entry, order status, real time portfolio, high/low scan, top gainers/losers scan, watchlist containing, in addition to my trading candidates, the major indices and technical indicators like VIX and TICK. Last but not least, is a chart window for the most promising candidate.

Discovering that ININ---set to gap up because it has given higher earnings' guidance---has a very small float and moderately high short interest, coupled with my knowledge that such stocks often move very strongly on low volume days, like Fridays, pre-holidays, and summer days, makes ININ the star attraction at this stage.

The market opens, and since I am particularly interested in seeing if ININ can trigger a trade, it gets top attention with a one-minute chart (above) on my main trading page. Meanwhile, I am flipping through my chart pages to see if anything is close to triggering---nothing yet. I check the one-minute chart of ININ, and I see the choppy pennant being formed at 10 A.M. The pennant range is around .25. That's enough of a sign for me to buy half of my lot as close to the bottom of this formation at 13.65. Buying a partial lot before the actual breakout happens makes sense in high-probability trades, especially if the desired lot size is substantial. I am set to buy the rest via an automatic buy stop if it can clear the top of the pennant at 13.87 on the 5 minute (see below). Trades moving strongly in the morning require monitoring via shorter time frames.

Fifteen minutes later, it does, and my full lot is in, average cost around 13.78. Just before the breakout, the pennant was around seventy-five percent completed and the volume dried up. Triangles/pennants need to be not completely formed or else they usually just fizzle out and do not support strong breakouts. The pause in volume is what my trading buddy calls a lullipop, a lull before the breakout, where the equilibrium between buyers and sellers end and the buyers take over.

Part of my prepurchase research of ININ was its support and resistance based on weekly price level resistance. Since the open of the July 24, 2006 weekly, bearish, high-volume candle was 14.81 (see below), therefore trapping lots of traders at that level, I determined the resistance to be around 14.80. The target of 14.80 also matches the length of the pole on which the flag formed on Friday's intraday charts, forming a measured move and confirming the resistance.

As the ININ trade continues, I am monitoring the price action on different timeframes, and have my mental stop under the flag for the first half lot at 13.53 and then just above the flag at 13.89 for the complete lot. Price is smoothly rising---each candle opens near but not with much overlapping to the close of the previous candle. I put my sell limit in for the full lot just under the target price. My offer gets lifted fairly easily, and I am out with a $1.00 move gotten in about 40 minutes. Since the risk was .25 for the complete lot, it was a +4R trade. For quickly moving stocks, I will often forgo hard stops, but it is a risk because my net connection could be interrupted. Then my carefully constructed trade will exist only in cyberspace floating without my pilotage. So, I am working on the artful science of consistently setting hard regular/trailing stops with upper triggers via bracketed orders to take care of that possibility.

My market basket is now empty until the next fresh thing jumps into it.

There seems to be a lot of controversial over the concept of R-Multiples. I've been seeing people complain about them for months now and I've been meaning to write a post about "R". I really wanted to do it last week but I'm glad I didn't get to it because this week a raging debate about R has popped up. Glenn, at DehTrader can serve as the poster child for the anti-R crew. Here's part of his recent rant against R-Multiples (emphasis is mine):

I post real numbers as opposed to R values, I always have. I like real numbers, I understand real numbers and I see truth in real numbers and I think the reader does too. As a reader of many blogs I find zero value in any post or summaries containing R values, I don't see any point in sharing that information. I suppose if I posted in R values I could look like a pretty good trader, but we all know I am a struggling trader. R can mean anything so why even bother with it... R stands for bullsh!t imo and that's my rant (that and ads haha). The best blogs out there post real numbers, Boogtser, JC (NYSE), the Kirkster all come to mind.

He's joined by folks like Paul who left this comment over on Ugly's post about R multiples:

I believe dollar values are more important than R value. I agree that the actual $ value is meaningless. However, R values are subjective and don’t give you a true idea on how successful the trade was. If you defined your risk at 15 cents and made 30 cents on the trade, while another person made 50 cents but decided his risk would be 50 cents, R values would say the guy who made 30 cents was more successful. I have a problem with that. It could very well be that the guy who only risked 15 cents is playing it too safe and his 2R gain was a bad trade.

So that gives you an idea of the anti-R sentiment. I'm going to explain why I think R-Multiples are so useful and why I use them in my trading and on this site.

What is R?

R is simply the dollar risk per trade. It's nothing but a reward-to-risk ratio. I first heard it called "R" in Van Tharp's book "Trade Your Way to Financial Freedom". In another of his books, "Financial Freedom Through Electronic Day Trading", Dr. Tharp reveals the great secret of trading:

The golden rule of trading is to keep losses at a level of 1 R as often as possible and to make profits that are high-R multiples.

You often hear (read) that traders should only look for trades with a reward/risk ratio of at least 2 or 3 to 1. Expressing your results in terms of how many times your risk allows you to easily see how well your trades measure up to such a standard. So when I look at my results in terms of multiples of R I can easily tell how good or bad the trades were. I like to think of R-Multiples as telling you the efficiency of your system.

So why not just use dollars?

Expressing my results in dollars would achieve the same result if I always risked the same amount of money. But what if I triple my account and therefore trade larger positions compared to when I started trading? Or what if I hit a rough spot and decide to cut my share size down while I ride out the storm? Then the dollar results won't easily tell me how trades from one period of time compared to another period of time. But if I use R making such comparisons is simple. Either my trades passed the risk / reward ratio test or they didn't. The actual number of dollars at risk doesn't matter, how many multiples of the dollars at risk does.

Along the same lines, recording trades in terms of R-Multiples allows you to easily calculate your system's expectancy. (Follow the link for why you should care about expectancy.)

Also, as Rx said:

talking and thinking in terms of R-multiple when you discuss about profits is an excellent approach - that by itself makes you focus on risk and money management - the actual "grail" to successful trading.

That is a very important point. Whenever I see people posting dollar returns, especially losses, that are all over the place the first thing I ask myself is "I wonder what his risk per trade is". It's almost a certainty that those traders aren't focusing on risk and as a result keep having huge losses. The mere fact that you have to define R and then place a stop to keep your loss to 1R is probably too constraining for those gamblers traders. Dr. Tharp says about determining your initial stop-loss point as soon as you enter a trade, which, by definition woud give you a 1 R loss:

This principle is so important that if you cannot follow it, then you might as well give up the idea of electronic day trading right away.

The reason I use R on the blog is because I don't want to discuss dollars or my account size on the site (as Ugly stated). That's nobody's business but mine. Also, it makes it easy for people to figure out what they could have made or lost on a trade with their own account size and risk per trade amount. If you see a trade that returned 3R all one has to do is plug in their dollar risk per trade to figure out what they could have made / lost.

To the R Haters

Let me address the "alleged" issues which I quoted above...

Glenn thinks that R is just some made up number and could mean anything. He likes "real" numbers. While it may be fun to see that somebody made $10,000 on a trade that in and of itself doesn't tell you how good that trade was. What if that person risked $30,000 to make that $10,000? Or what if they risked $1,000 to make that $10,000? Those are two very different trades. Sure they both made the same amount of money but isn't the second trade a much more efficient use of capital?

What if somebody is trading $500,000 lots to make $1,000 in profit? It may be nice to see somebody saying that they made $1,000 here and $1,000 there but damn(!) that's an inefficient use of capital. So while R could mean anything in terms of dollars, in my humble opinion what really matters is how many multiples of R were made or lost. That tells you the quality of a trade or system.

Glenn also states that if he reported his trades in terms of R he could appear to be a good trader. I'm sorry to tell him that's simply not the case. If you lost money that means your expectancy, which is just your average return expressed in R-Multiples, was negative.

Paul said that "R values are subjective and don’t give you a true idea on how successful the trade was". That is exactly wrong. R-multiples are the very thing that tells you exactly how successful a given trade was, if you choose to grade on a risk/reward basis.

Percentages vs. Dollars

This debate about R reminds me of a conversation I had a couple of weeks ago. I was in a presentation for Trade-Ideas' new tool, the Odds Maker. They were showing how you could backtest all these different scenarios with the tool. The results were expressed in average dollars won or lost. Another viewer and I asked about seeing the results in percentages. They kept saying that perhaps they would do that in a later revision. I kept harping on it because to me seeing the results in dollars was of little use for the way I size my positions

The argument from the presenter was that all you had to do was multiply the average dollar return by your average lot size to figure out how much money you could have made with a given system you were testing. I had to disagree because my lot size can vary drastically depending on how far away my stop loss is. Here's a situation which could be problematic -- I trade Google with a 2 point stop (which is only about half of a percent) and get lucky and make 6 points of profit. All of my other trades are on stocks under $50 with stops less than 50 cents. I could have some combination of winners and losers mixed in there... most of them probably well less than $6. That $6 gain may skew the results when presented as average dollars won. That's an over-simplification and there are all kinds of possible permutations. But I hope my point is clear that looking at the results in terms of average dollars won/lost may not tell accurately tell you the story.

So how can we make the results clearer? Simple, express them in percentages. That way, regardless of how many shares were traded or the prices of the stocks traded the results can be equalized across all the trades. I feel much better being able to say , "OK, this system would have returned X%" instead of "X number of points.

We debated the merits of each way of reporting for a few minutes and at one point somebody said, well , for this release we're aiming for the "lowest common denominator". In other words, the average person can't think in percentages, so we're just gonna report in points. I was like, F the average person, make it work the "right" way! The funny thing is that after debating all of that the software actually could express the results in percentage terms. We just had to switch a setting.

So my point of that little story is that I always prefer to think in terms of percentages in stead of points. I always see people talking about number of shares of point moves. For example, you might hear somebody exclaim "Google is up 5 points!!!" I don't see that as anything to get excited about. That just over a 1% move -- a normal fluctuation. You'll hear similar things from reporters talking excitedly about the Dow being up some triple-digit amount. The Nasdaq may actually be up a lot more on a percentage basis but they'll just say, eh, the Nasdaq is "just" up 30 points.

Looking at the percentages makes those kind of comparisons easier. R-Multiples do the same thing for traders. They can accurately compare their own trades and they can take another trader's results expressed in R and easily relate them to their own system.

Cameron just left a question for me in the comments to last night's recap. He asked:

I'm looking at your way of thinking about the market. One of the biggest questions I have, though, is regarding your stops.

You said you set your stop at just .20 over your short dollar amount. With all of the intra-day fluctuations, how can you know that it wouldn't have gone up, say, .25 and then nose-dived? It's a perfectly plausible situation.

Said another way, I can place a buy order with a stop, but we easily have 1% swings in a day. If a swing occurs, I have automatically sold for a lower price, only to see the market go back up. Therefore, I'm encouraged to set much broader stops. You set yours for only .20 over the buy price, though, which is why I'm confused.

I often get asked variations of this question, so I thought I'd put my answer in its own post. This was my response:

I couldn't have known that the stock wouldn't have shaken me out. But I also wouldn't have known if I'd have been shaken out if my stop was 50 cents or even a dollar higher. That narrow range candle just provided me with a tight stop. I like tight stops b/c they give me a better opportunity to get high R-multiple winners.

But to your point, how tight of a stop is too tight and will practically guarantee being stopped out? It's hard to answer that question but I generally shy away from any stop that's less than 0.5% of the stock price. I normally like my stops to be about 1% of the stock price. In general, I'm more comfortable with somewhat tighter stops on the QQQQ b/c of how liquid it is and b/c it's less volatile than an "in play" stock would be.

Also, note that I didn't pick a 20 cent stop out of thin air. I went with what the previous candle told me was resistance.

The key thing to remember is that I'm on the hunt for high R-Multiple winners. Because of the way I size my positions (the percent risk model), the tighter my stops the higher the potential reward. In other words if I think a stock has potential to move $2 I'd rather have a 20 cent stop than a $1 stop. Risking 20 cents to make 200 cents ($2) would give me a 10R winner vs only a 2R winner if I had risked $1. (If R was 1% of my equity that would be a 10% gain vs a 2% gain.) For a real example of this see the DGX trades I posted on my Narrow Range Bars post.

So there' a trade-off at work here between stop size and risk/reward multiples. The tighter the stop, the better chance you have of getting a high R winner but you also have a better chance of getting stopped out. Your win ratio is likely to be inversely proportional to how tight your stops are -- looser stops should lead to a higher win rate but also fewer high R-multiple winners.

It's also key that you don't pick your stops willy-nilly. I always use the support/resistance levels that the charts dictate to me.

There's some good advice in Alan Farley's latest article to help "home gamers" (Cramer's phrase) survive the current extra-volatile market. The whole article is worth reading but here are the four bits of advice:

I have four pieces of cautionary advice for my at-home brethren. Listen up, because it's no longer a question of whether you want to take real money out of the market, or just add a few bucks to the weekly shopping budget. These remedial steps must be taken if you want to survive long enough to take advantage of the real opportunity.

1. Wait for the Market Volatility Index (VIX) to drop below 40 and stay there for a week. Massive price swings require equally massive stop losses, which rarely justify the intended positions. Your only alternative is to stand aside and do nothing, no matter how much it hurts to watch others playing those big rallies and selloffs.
2. Forget overnight positions until the index futures stop gapping 2% or more every morning. These price jolts are great news when you're on the right side of the trade, but total devastation if you're on the wrong side. And guess what? You're not smart enough to predict overnight direction from day to day. Neither am I.
3. End your love affair with popular stocks that made you money during the last bull market. In November 2008, these are the issues that will trigger the most painful and unexpected reversals, which happen right after you're absolutely convinced your position is the right play. The bottom line: They see you coming, sucker.
4. Get control of the time element in your market strategy. You're getting killed because you have no patience and forgot how to sit on your hands when your trading edge isn't in play. Realistically, it could be months before the market works for you again. Would you rather wait it out and survive, or stay busy and get crushed?

My thoughts:

1. I like that part about watching the VIX. A few weeks ago I decided to cut my risk per trade (R) in half (from 1% to 0.5% of my equity) until the VIX breaks that top it's trying to put in. That means a close below 44. (Dr. Brett recently wrote about when traders should cut their risk.)
2. Tough to argue with what he said about overnight holds. This is certainly a market for intraday trades.
3. Reminds me of what I said about AAPL and GOOG back in July. Add DRYS and its ilk to the list too. It's better to just have an open mind and grab on to whichever stocks become the new leaders.
4. Yep, preservation of capital must be the top priority. If you can't sit out at least lower your risk.

I was recently asked about how I determine my "universe" of stocks. That is the stocks that I consider fair game to be traded. I use different filters for swing trading and day trading but they do have similar criteria. In both cases I'm trying to find liquid, volatile stocks. Once the filter's applied I then look for certain patterns / setups to trade.

For swing trade candidates I use the following criteria in TeleChart:

• Price Per Share: about $5 to Max
• 5 Day Average Volume: 75th percentile & higher (right now that works out to about 466,000 shares per day)
• Price Volatility: 25th percentile & higher
• Volume (Dollars) 5-Day: 75th percentile & higher

Here's a shot of the filter. The count column shows how many stocks are left after applying each row in the filter.

That typically narrows the market down to about 1,000 stocks. Then I look for setups among that set of stocks. I gave a lot of details about the setups I scan for back in the day.

For day trading I'm trying to find liquid stocks which are abnormally active and volatile. To do that I use the following two scans in Trade-Ideas:

Bearish Scanner's Filter:

• Min Price: $10
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: -0.5% (negative means down at least 0.5%)

Bullish Scanner's Filter:

• Min Price: $5
• Max Spread: 10 cents
• Min Current Volume: 3 (stock is on pace to trade 3 times its normal volume today)
• Min Volume Today: 300,000
• Max Up from the Close: 1%

Sometimes I'll tweak these settings to give more or fewer alerts. I usually do that by adjusting the "Min Current Volume". Details about the actual alerts & setups I track for daytrading are on my "how I work" page. If you use Trade-Ideas you can paste the following link into the "Collaborate" feature and load that filter and all the alerts I track:

Bearish -- grab the URL of this link

Bullish -- grab the URL of this link

One thing I have always preached on the blog, less so on Twitter, where I bang out more trading ideas and market thoughts is that shorting stocks is hard. I think it’s harder than any aspect of learning the market. It’s dangerous. Let this morning be the only reminder you should EVER need.

Howard's right about shorting being difficult. I think it's so hard because it's not simply the opposite of going long. (I won't even go into the whole thing about your losses while short being theoretically infinite. Been there... done that. Use a stop to limit your losses, use proper position sizing, ONLY short LIQUID stocks and you'll be fine.) What makes shorting tricky is that bear moves often have violent (short-covering) rallies because the psychology of the crowd trading a down market is different than that of a bull market. You have to be quick on your feet when shorting. My motto is "stick & move".

Many traders love to buy breakouts in bull markets. (Whether that's actually a good strategy is debatable). In my experience swing trading, the opposite of that strategy, shorting breakdowns through resistance, will often lead you right into a snapback rally and, as MaoXian used to say "the quickest loss ever". That's why I often make note of all the people who are initiating shorts after the market has already fallen to a major support level. We saw that this week as I noted in some of the morning watchlists.

My contention is that if you were caught short Friday morning you should consider your losses as tuition paid to the school of hard knocks. Learn from that expensive lesson, take your losses and hopefully survive to trade another day. I'll stop short of saying that the losses were deserved but there were plenty of warnings to at least cover your shorts if not to get long. There are so many good sources of market information these days, both on the web and, yes, even on TV. I'm not saying to blindly follow somebody else's opinion but it can be helpful to see what others are doing based on what they see. Here are just a few of the recent warning signs:

• T2108 dropping below 20 -- Ah, good old reliable T2108. I've been watching it closely as we've sold off. On Wednesday I noted that it finally hit the point where wise shorts would want to cover. It pays to find a good overbought/oversold indicator and heed its warnings. (You may need different indicators or settings for different timeframes.) Sell at overbought and cover at oversold. A couple of years ago I decided to force myself to put some IRA money to work every time T2108 broke 20. It hasn't failed me yet.
• The Fed (Plunge Protection Team) has interfered announced stimulus packages around the July lows a couple times. On Thursday and earlier in the week there was talk of more PPT action.
• The PPT took action earlier in the week and last week. Just look at the orchestration of the LEH and AIG situations, etc.
• Extreme Volatility -- The moves all week were nothing short of violent. Positions were whipsawed all over the place. That in & of itself would be reason enough to lighten up on positions if not move to the sidelines. That kind of volatility is often a sign of a trend reversal, not of a continuation of the previous trend. That's why so many traders watch the VIX. Tons of people noted the spike in the VIX on Thursday.
• Corey from the 'Afraid to Trade' blog warned 'Use Extreme Caution in the Week Ahead.' He gave many good reasons, including the Federal Reserve interest rate decision, the quadruple witching options expiration and headline risk from troubled financial firms. His crystal ball was working well when he wrote "We could see a week ahead that will be discussed years later - as such, if you are a newer trader, it might be best to switch to simulation mode this week or use this week as a training experience, rather than risking real capital in an environment that could swing violently up and down due to market events scheduled to happen this week."
• Bullish Technical Divergences -- Dr. Brett pointed out some bullish divergences he was seeing in the market. Perhaps most important were what he called 'those fuzzy indicators' -- "Traffic on the blog is way up, reflecting trader uncertainty and desire for information. I just fielded my fourth media interview request in two days. During quiet and bullish market periods, I don't get four requests in a month." On Tuesday morning I also noted my blog's traffic spiked as did Barry Ritholtz. I've often joked about making some kind of sentiment indicator based on my site's traffic ebb & flow & referral logs. It's not a bad idea and I think it would be especially useful if it were based on a major financial site's traffic data.
• Dennis Gartman telling folks to "be small" -- Gartman was on CNBC's 'Fast Money' early in the week saying that he was scared of the market's movement and he was "being small" and planned to "get smaller". His advice to others was to "be small" in this market.
• Jeff Macke, also on 'Fast Money' was warning people not to play (trade) if they didn't understand the (changing) rules of the game. He was referring to all the headline risk from PPT action and the volatility caused by rumors -- many of which were spread by CNBC during the trading sessions. Ironically, Macke was kicking himself Thursday night for not following his own advice and getting caught short.

I fully believe that had the PPT not acted on Thursday night the market was *destined* to move higher in the short term on its own. Still being short on expiration Friday in this environment was just asking for trouble. So what's a trader to do? Like I said earlier, stick & move. I think it makes much more sense to short bounces back to a trendline or moving average. William O'Neil's book on short selling talks about using retracements to the 50 and/or 200-day moving average as a place to initiate shorts. By the time the stock (or whatever instrument you're trading) is extended from its moving average it's time to cover. Then you can wait for another bounce to reload.

The strategy to cover & reload makes more sense when you're short due to the fact that if you're dead right with your call on the stock dropping the most you can gain is 100%. The odds of that happening are slim and if it ever did happen you'd likely have to ride out some severe short squeezes. But you might be able to stick & move your way to more that a 100% gain by reloading multiple times. In theory, you could catch 15 10% moves lower in a stock that's falling, retracing & falling anew.

Short sellers need to be nimble, pick their entry and exit spots wisely and heed signs of impending reversals. Trying to short breakdowns of an already extended market is a sucker's play -- as is overstaying your welcome while short.

I got this question the other day:

When you short a stock at 30.00 and you want to place a stop limit in case it jumps up past your max loss of say, 31.00, how exactly is that order entered (everytime I try to do this, it sells (Ed: he means buys) my stock immediately - I'm doing something really wrong). Do I put a limit price of 31.00 and a stop price above that, i.e. 31.50 (thus saying it's ok to repurchase any shares between 31.00 and 31.50)?

This is one of the trickier orders to place. In that situation you would want to enter a buy stop limit order (buy on stop with a limit). The stop would be at 31.00 and the limit would be 31.50. Here's a screen shot of a similar order to cover QQQQ with a buy stop @ 47 with a limit of 47.10:

What would happen here is that once the stop price of 47.00 is hit a buy limit order of 47.10 would be entered. As with any limit order, you're guaranteed price but not execution, so you may not be filled. If you wanted to be sure that you got out (covered) just use a buy stop of $47.00.

In case you missed it in the links section the other day, be sure to read the interview with Michelle B over on the StockTickr blog.

The other day Downtown Trader asked me how I handled trailing my stop on that DGX short. I pretty much followed exactly what's in my post about how I take partial profits. I moved the stop to breakeven once my gain was equal to my initial risk of 50 cents per share. As it dropped further I moved the stop to lock in one-third of my profit. Why one-third? It's a value I came up with which I think allows a stock to have a normal retracement without stopping me out. It's sort of an arbitrary level, although I got the idea from the Fibonacci levels. I've found that method to work reasonably well after much trial and error with trailing stops. I read somewhere that you can consider a retracement to be normal as long as it didn't break the 62% Fib level. So I figured that I could trail my stops somewhere around that level to allow stocks to have their normal fluctuations. I generally get more aggressive with the trailing stops as the day wears on. I'll move them to 50% by 3:00 and just keep pushing them until the close -- assuming I don't bail before then like I did with DGX.

The other day I got the following email about narrow range bars:

I was going through your site and others and I keep on hearing, no entry on narrow range bars. What is a narrow range bar, and could you send me a picture of one?

A narrow range bar (candle) is simply a bar which has a range from high to low that's much less than the average bar for a given equity. You'll often see people, like Dr. Sen, tracking NR7 bars. NR7 means the narrowest range of the last seven bars. For example, here are some NR7s from last Friday:

So now that you know what narrow range candles are, why should you care about them? Since stocks generally cycle between periods of low volatility and high volatility it can be beneficial to identify when the high volatility times are approaching. Often times, especially for trending charts, narrow range periods -- periods of range contraction -- will presage range expansion. You can think of a stocks movement like a spring / slinky -- it need to recoil (contract) in order to build up enough potential energy for its next expansion. However, it's important thing to note that range contraction doesn't tell you the direction of the impending expansion only that an expansion may be coming. It's up to the trader to use other means (indicators, common sense?) to catch the expansion in the right direction.

One of the nice things about trading against narrow range bars is that they give you a tight stop. If you size your positions like I do, the smaller the stop means you can buy (or short) more shares. It's the trades with the really tight stops that let you make those huge R-Multiple trades that often make or break a trader. Here's an example from a trade I took last week.

I'm going to use the actual entries taken by myself and by Butterboy but instead of using my actual exits, I'll assume I held until the close. So let's say my equity is $100,000 and my R, the amount I want to risk per trade, is 1% or $1,000. Given that, here's the first trade, based off of my actual entry parameters:

• Entry at 53.96
• Stop was 50 cents higher so I could short 1000/.50 or 2,000 shares
• DGX closed at 49.64, which is $4.32 lower. So that trade returned $8,640 or 8.64 times my risk (8.64R)

Here's the DGX trade using Butterboy's entry criteria. He used a 15 minute chart and entered lower than I did but his stop was also much tighter:

• Entry at 53.40
• I believe his stop was 21 cents higher, just above the 11:30 bar which he entered against. Given that one could short 1,000/.21 or just over 4,700 shares.
• DGX closed at $3.76 beneath the entry so that trade returned $17,672 or 17.67 times the initial risk (17.64R).

It should be clear why so many of us like narrow range bars and why I scan for them. That second trade had a worse (lower) entry but because of its tighter stop it was almost twice as profitable. It also shows why I love the percent risk position sizing model.

Here's