Holding: EASY

Position Sizing: how to determine the number of shares to buy. A common approach to determining the number of shares to buy is to figure out how “good” the opportunity is—the better it is, the more you should buy—(not a good idea!), or perhaps to set a fixed dollar amount per position. This last approach at least limits the risk in some way, but is far from ideal. This is because, although it balances cost, it does not balance risk, which is not the same thing. Using computer simulations of various trading systems, Van Tharp found that profitability can be enhanced many times over, using the same buy and sell signals (that is, without improving on the actual prices bought or sold, and without regard to the particular stock selection process), by basing the number of shares purchased on the risk level selected, rather than on a fixed dollar amount. This matter of “selecting” the risk means that you always know before the purchase is made, approximately how much you are going to lose, in dollars, if you’re wrong—by pinpointing where you are going to get out (using a stop loss order) in the “unlikely” event that the trade heads south. Note: a good 50% of all trades head south.

Although I actually figure out what I’m going to buy, and where I’m going to buy it, before I decide where I’m going to sell it, I also always figure out where I’m going to sell if I’m wrong, before I place an order. I mention this here because in principle, if you can learn how to control (limit) risk, it’s much easier to succeed. You can never know beforehand how much you might make if a trade is profitable, but you can always know with certainty where you’re going to draw the line if you’re wrong (you may of course be filled somewhat below that line). This is what’s in your power to control. Furthermore, it’s far easier to make this decision before you take hold of the hot potato. Oddly, for most traders, the fear of losing a profit is greater than the fear of not having a losing position recover. Ideally, fear of having a small loss become a big loss should be so intense that it would be unthinkable ever to enter a trade without a predetermined loss limit, or to hold onto a trade that has reached one’s limit. In order to control risk, as well as to maximize profits, it’s also essential to know how to decide how many shares to buy. Almost nothing in print even discusses this. Van Tharp is one of the few authors that does. Here’s how to determine the number of shares to buy (1):
a. Limit your maximum risk to 2% of total equity. Once you’ve set that as an upper limit, you’re free to modify it based on your own personal tolerance for risk. As my boss used to put it, when opening an account for a new customer: “How much do you want to lose?” This matter is especially critical for anyone that is young, old, educated, uneducated, successful, intelligent, or a stereotypical male. The reason is that they assume that the behaviors behind their successes in life so far are either a result of being aggressive and courageous when it comes to taking risks, or the result of using their brains. And they wrongly conclude that their intelligence will pretty much assure the same degree of success in picking winners that it assured in school or in their careers. Being wrong isn’t an option. This is why almost all such traders lose. In school—in fact, in just about everything in life– we’re rewarded for how often we’re right. In the markets we’re rewarded for how consistently we can be wrong small (even if often), and right big (even if less often), which is an entirely different matter. Remember, the balance in your checkbook is the sum of the deposits, less the sum of the withdrawals, not the number of deposits or withdrawals. Successful traders apply their intelligence to objective criteria, and focus on limiting losses and letting profits run, rather than on prediction. When it comes to trading, it’s smarter to exit losing trades quickly, rather than hold onto them forever (“it’s just a flesh wound”) while you bleed to death. If you’re inclined to savor the pain, it’ll be more profitable taking many small losses than locking down and permitting any single loss to become a big one.
b. Where risk is concerned, individuals vary greatly in their tolerance for pain and stress, but probability needs to have the final say in this matter, not one’s ego or emotional ability to bear loss. You may personally be inclined to risk 20%, 50%, or all of your capital on a given trade, if it “looks good enough,” or because it’s the macho thing to do. But this is where the research is clear: if you systematically risk over 2% of your total equity on your positions, it’s only a matter of time before your losses are so great that you will never recover. Your courage, intelligence, or tolerance for pain (“no pain no gain”) will not save you. If you are fairly conservative, as I am, you may decide to risk less, but risking more than 2% greatly increases your risk of ruin. My personal decision in this matter—based on an extensive independent review of my actual profits and losses–is to set a maximum risk for each trade of 0.006% (6/10 of 1 %) of total account equity. At the same time, I want to allow for the occasional catastrophe in which my ideal risk limit of 0.006 % is greatly exceeded by adverse market action. And since I have only observed a few instances out of 1000’s in the last three years in which stocks that meet my criteria have opened over 50% lower overnight (in which case my sell stop would be activated at the market), I have reasoned that if I allocate up to 4% of my total equity for a trade, then in those rare instances in which my position opens as much as 50% lower, although my 0.006% risk will be exceeded, I will come close to limiting my total loss to 2% of total equity (50% of 4% is 2%). I set an upper limit of 4% of total equity for the amount allocated to any position, but also place a stop with a maximum targeted loss of 0.006% of my total equity. If the trade goes against me, my stop will limit my loss to about 0.006% in almost all circumstances. In practice, allowing for a bit of “slippage,” my largest single loss has been 2% of total equity; the second largest is 2/3 of one percent, and my average loss is about 1/3 of one percent (less than the targeted 0.006%, because in some losing trades, my stop advances somewhat before I’m stopped out). You could justify selecting even less risk than I have selected, with safer but smaller long-term results. Whatever your decision on the level of risk to take on, just make sure that you maintain the same percentage for all of your new trades at any given point in time. If you want to change it, change it for all of them. As the total account equity goes up or down, maintain the same risk limit for each new trade. A special caveat: if you have traded relatively little, the 2% maximum may sound extremely low (it did to me), and you might be inclined to just throw this whole piece out the window, and put some serious money on your most promising trades. I’ve heard professional fund managers self-righteously talk about doing just that with their “best” ideas. What you won’t hear is how they fared when their best ideas don’t work out. Imagine putting what might sound like a fairly reasonable 10% of your account on each of ten difference positions. It’s extremely common to have 10 consecutive losses, not to mention a few dozen. By my calculations, ten times minus10% won’t leave much to trade with. Unless you keep your losses ultra small, you will eventually reach a point of no return. Trust me, the draw down possible at just 2% per loss can be quite numbing. It’s still your call, but the more I live with my 0.006% targeted loss limit, the more I like it. A recent refinement: since I use a trend following system, which by its nature does worse in choppy sideways markets, and better in strongly trending markets, I have recently made a modest change to my risk limit. If I’m experiencing a drawdown, my risk limit gradually drops by small decrements. And in a market that’s beginning to trend strongly, it gradually rises. This means I lose less than I otherwise would in draw downs, and take more out of trending markets.
c. Determine your risk per share on the stock in question: this is the difference between your buy point and your stop loss point. (Unless you place a limit order, you can only know the buy price approximately before you get the confirmation of your purchase; estimate by using the projected buy price + .01).
d. Determine your initial stop: this is generated automatically by my program, and equals the most recent intermediate low immediately preceding the buy point (2). In practice, this is usually the lowest low in the last few weeks, but never significantly below the 50-day average. If it’s closer, it’s likely to be inside the “noise” range, which means too close! I want to limit my loss, not make one virtually certain (which is what will happen if it’s too close). Before going further, I should add that basing your stop on chart points is not necessarily the best or only way. In fact, I strongly suspect that a stop based on a fixed number of standard deviations (say, two) away from entry point might work better. Given my extreme personal time constraints, and success using simple chart points (3), I continued using these through the end of 2003. However, since chart points also have some drawbacks, I have finally decided to take the extra time required to base my stops on a more sophisticated series of calculations, which I’ve previously tested and used extensively in commodities. So, starting in 2004, all of my stops have been based on this program.
If you are following the above steps to get in and out of trades, then it follows that several extremely common behaviors must be avoided, because they increase risk and minimize profits: These are A) Averaging Down, B) Exiting when a “target” has been reached, and C) Buying a stock because it has fallen so much that it’s now “cheap.”
A) Averaging down:
NEVER add to a position after your initial purchase. Although this is a common practice, and is recommended by IBD, among many others, whether averaging up or down, it clearly violates the rules on risk. The “reasoning” behind this practice is that if you buy more shares when the market is dropping, you’ll then make even more when it recovers. This is based on the fatal belief that stocks that go down are likely to go up (the opposite is true), and totally ignores the increased risk of ruin that comes with those extra shares. While a budgeting case can be made for regular (say, monthly) contributions to a retirement program, the use of that contribution to buy more shares of a particular stock or fund would only be valid if these mutual funds or stocks meet these criteria at that very moment. If this were to involve adding on to positions that are dropping or that have high value ratios, it would be a clear losing proposition. If you are making monthly contributions to a deferred comp program, one way to avoid committing this serious blunder is to funnel all deposits into the money market option of your program during down markets, and shift it back—100%– into stocks when the market signals that a new upmove is beginning. Regarding adding to losing positions, as newsletter writer Dennis Gartman puts it, this tops his list of 22 rules: “Never, under any circumstance, add to a losing position…ever!”(4)
B) Exiting when your target is reached:
NEVER exit because a stock has gone up as far as you think it should. Never set an upside target. Ignore the targets set by brokers. They are hype, pure and simple, designed with the sole purpose of trying to get you to act. There is serious evidence that Wall Street analysts’ earnings projections have no value. David Dreman in his important analysis of their research, found that they are overly optimistic in bull markets, and overly pessimistic in bear markets. They are way off the mark in both instances. “In the October 11, 1993 issue of Forbes magazine, Dreman recounts a study that used a sample of 67,375 analysts’ quarterly earnings forecast estimates for companies on the NYSE and AMEX between 1973 and 1990. It found that analysts’ average forecast error was 40 percent and that estimates were misleading by more than 10 percent two thirds of the time.”(5) This represents an enormous margin of error, since a change as small as 1% can cause a stock to soar or collapse.
In spite of my own advice, I have sometimes exited positions that are still in an up trend. I discovered how serious this problem was while compiling a detailed review of my trades for 2001. I saw a pattern in my trading that was disturbing (My name is Don, and I…). I was getting out of good positions on short term signs of weakness, such as classic key reversal days, or a series of new highs on decreasing volume, or sudden down days on very high volume. As valid as these “sell” signals may have appeared for the short term, they were not strong enough to invalidate the longer term up trend that had triggered my still valid positions. I watched in disgust as one stock after another resumed its up trend within days or a few weeks after I got out, without having dropped to my own original stop. I had fallen prey to the “fear of losing a profit.” By actual count, I left on the table profits over twice as large as the ones I’d realized. This has given me a certain amount of encouragement, regarding my future potential, because I have a clearer picture of what I need to do to reap bigger profits. In December 2002 I officially deleted all short term sell rules from my program, and it has paid off royally. The rest of the program remains intact. My only remaining sell rule is to exit when my stop is hit. I may not be well yet, but I’m doing much better.
There is no way to know what price a stock is going to go to. Get in and out when your rules tell you to.
C) Buying a stock because it’s “cheap.”
NEVER get into a stock that is in a down trend or because it seems “cheap.” The research on this is probably more conclusive than on almost any other matter. The fact that a given stock used to be much higher in no way means that it’s now a bargain. Last year’s styles on the bargain rack are there for a reason, and are not likely to appreciate. Stocks, at least as held by the average trader, unlike any good or commodity, have no inherent value. A company may declare bankruptcy, and owe the holders of its common stock nothing whatsoever, and continue in business. At best, a share of stock is not even an IOU. The stock of the company is not the company. The company can make good profits year after year, while the stock goes down or sideways. The stock is a separate entity with a life all its own. The vast majority of all stocks that are beaten down simply get cheaper, and throughout the history of the market, these often become worthless. The more beaten down a stock is, the more likely that it will never recover.
As mentioned above, it’s very common for traders to determine the number of shares to buy by simply dividing their total capital into equal piles of money. The one advantage of using equal dollar amounts is that it’s easy. It really has no other advantages! The advantage of basing trades on risk, instead of equal dollar amounts, is that when you’re right, you take more profit out of a good move, and when you’re wrong, you still systematically limit your risk and greatly minimize the volatility of your account. And knowing your risk in advance makes it infinitely easier to live with trading. Here’s an illustration of position sizing based on simplistic equal dollar amounts (top half of the table), compared with a position sizing approach based on risk (bottom half). Compare winning and losing trades in each version:

Note: In the above simulation, buy prices, initial stops, and sell prices are identical and there are the same number of profits and losses (5 each) in both versions. Total outlay for all trades are the same ($100,000). Yet, profits are about double in the second, risk-based method: $39,446, vs. $19,947. And the average risk is 27% higher ($1269 per trade) in the equal dollar approach, vs. a fixed risk of 1% ($1000 here) in the risk-based example.

Let’s see how position sizing works with some hypothetical stocks: OK, they’re real. [Note, 7/20/08]: In the current market, there are very few stocks that meet my criteria. Here are a few that do. Now, I haven’t the faintest idea as to whether these stocks are going to go up, down, or sideways, now or ever, but after the recent down move, as stops were hit on one position after another, I now have one third of the account in cash , and it’s burning a hole in my pocket. I’m going to buy any of these that break out on volume, starting Monday AM, 7/21/08.

The point of this chapter is to show how my position sizing works. I cover selection criteria more thoroughly in Chapter 12, but will briefly touch on the fields in this table.

Since I know from James P O’Shaughnessy’s long term studies in What Works on Wall Street, that stocks with a low PSR and high RS more frequently and more significantly outperform virtually any and all other stocks, I start by screening for all stocks with a PSR under 1.0, and a high RS. I can’t buy all 213 that showed up, so I then eyeball the charts for all of these and look for any that have been in a sideways pattern for a number of weeks. Yes, that’s admittedly a highly subjective process. Regarding the number of weeks in the pattern, IBD recommends a minimum of seven weeks, but has noted that fewer—as few as three–may be sufficient for smaller stocks. I’ve tracked this question for eight years now, and have found a strong correlation between the number of weeks and the profitability of the trade: the more weeks the better. You’ll see a few stocks on the list in five-week patterns. My long term average is about 13, so I can stand throwing in a few shorter term patterns.

Regarding Capitalization, various studies, including O’Shaughnessy’s, have found that small caps outperform large caps long term. But O’Shaughnessy himself notes that if you throw out the smallest and least liquid of these, the small cap effect is greatly reduced, or eliminated. That being said, he has broken out the performance of all stocks by capitalization, from smallest to largest, year by year, and his own numbers strongly confirm the belief that small outperforms large, long term. The difference between any two adjacent categories is significant, and the difference between the largest and smallest (after throwing out all of those that O’Shaughnessy found too small to even consider) is huge. But it’s also true that from time to time, large caps outperform, so I allow for that in my own selection process. If all of the highest rated stocks with low PSR and high RS above were large caps, that would then tell me that they are the ones to buy.

The same comments might be made regarding value ratios vs. growth criteria. Sure, sometimes value rules, and sometimes growth. But this is another non-issue, since a stock with a low value ratio can also be a growth stock. And common wisdom holds that earnings are key, but do you see any earnings ratios in my basic selection criteria? Not a chance. O’Shaughnessy found that stocks with the highest rate of earnings gains over the last one or five years significantly underperformed almost all other criteria.

And although one might factor in a great many other fundamentals and technical indicators, at a certain point, I draw a line. I just take a final look at IBD’s and Navellier’s overall scores, and narrow the list a bit more. The stocks above made my final cut.

But wait, I forgot to mention what all of these companies do. I didn’t mention their products, the story behind each one, and I completely forgot to factor in Wall Street’s earnings forecasts, recommendations, and price targets. That’s because none of that is in any way relevant.

I then download these into my trend following program. This then tells me where my initial stop will be, and once I know that, since I can already see my hypothetical buy point, which I found by examining the chart, I now know how many points I’ll need to risk initially per share. Then, plugging those numbers into my position sizing formula, this tells me exactly how much I’m going to risk, how many shares, and how much it’s going to cost.

At this point, you might wonder, since I know that these are all highly rated, why not just take the plunge with all of them right now? Well, that’s an option. O’Shaughnessy’s work clearly shows that stocks that meet these minimum criteria could be expected to outperform. But I also know from my own research that a couple of other ingredients are very important: volume and range. High volume and high range are both correlated with a further move in the same direction. So, if I wait for a high volume high range breakout on these, I’m going to do better than just buying a basket of highly rated stocks.

In case you’re tempted to evaluate any of these candidates based on what you think their potential might be vs. any others that I might mention with ultra high PSR’s or low RS, you’ll be making a serious mistake. Put a black mark by your name and reread David Faust’s notes on the limits of human judgment (see Bibliography). Here’s what this portion of my position sizing sheet looks like after plugging these candidates in:

At this point, these are just potential buys. They must now make high volume, high range breakouts. After lengthy reviews of the volume on my 3500+ trades for the last eight years, I’ve concluded that the definition of “high” must be much higher than we’re commonly given to believe. IBD suggests that 50-100% is high. My own trade history shows that something closer to a 400% increase is needed to show a strong correlation with profitability. But since volatility varies greatly from stock to stock, I’ve also found that basing my definition on a Standard Deviation is more reliable, with Two StDev’s being highly significant. In terms of range, if you have high volume without a significant change in price, then, by definition, all you really have is churning. I want to see high range too, and have found that setting a minimum of 1 StDev for the range works great. And since I want to know in advance when those volume and range limits have been reached, I figure that out in advance, and place an alert on my brokerage website so that I’ll be able to know when it happens.

STOPS:

Assuming that one or another of these trades meet these final requirements of high volume and range on the breakout, I place my order, and then as soon as I’m filled, I enter my protective sell stop, good till cancelled. The fact that a stock meets my minimum criteria is no guarantee that it will be profitable, but if it is, based on my program’s history, the size of the profit is likely to be several times the size of the potential loss. That’s why I’m going to buy it, not because of its projected earnings, or what some so-called professional analyst thinks, or any other similarly irrelevant piece of information. Remember, you must have a stop, because a good half of all trades fail. And if and only if they work, you must give good trades time and room to keep working. Profits take time. If you violate either of these principles, you can lose money on what otherwise might have been excellent trades.

Many approaches to placing stops have been proposed, based on a fixed percentage of the price of the stock or of the dollars put at risk, or simply a fixed dollar amount. Some of these approaches are even recommended by noted “authorities.” Although these sound suspiciously similar to basing risk on a percentage of total equity, they are really quite different, and simply do not hold up to scrutiny. William O’Neil’s rule is always to limit your loss to 7 or 8% of the purchase price. Well, fluctuations of this magnitude are virtually a daily occurrence, which means that this is far too close. As it turns out, the stops on the potential buys above start at 8% and go all the way to 20% below the purchase price, and those are based on the most recent short term lows, which is the closest that you can place a stop without being deep within the noise range. It’s not the distance from your buy point that matters. It’s the risk to the total equity in your account.

There is research that shows that giving a stock more room to breathe initially (but less as the profit accelerates) leads to greater profits in the long run. This is the path I have chosen.

What about exiting at a predetermined upside target? Well, yes, it’s possible to calculate fair value. And yes, there are such things as key overhead chart points for some trades. If that’s where you want to exit, you’re going to need to be right often and big to make money. Since there is no valid theory for projecting whether a given trade will work, or how far it might go, I make no effort to set an upside exit point. TARO, a couple of years ago, starting from a few dollars per share, went to about $100. TASR went from $4 to a pre-split price of $203. Then there were those legendary moves in the 90’s, as with QCOM.

However, I do know that on average, if a given trade works, I’ll make about 2.5 times what I’ll lose if it doesn’t. And since there is no theory that can tell me that this trade is even a bit more or less likely to work than not, I will make the trade if it breaks out on the upside with the required volume, and I will reject the trade if it doesn’t. I can only acquire a false sense of security from thinking that I have a better than 50/50 chance of knowing the future of this trade. And since no such knowledge is possible, my only recourse is to think of it in a probabilistic sense, as one of many trades, which are subject to probabilistic statements, and assure myself a real sense of security by entering a stop in case this trade fails. If the stock does break out on the upside, after I’m in it and after I’ve entered my initial stop, instead of thinking about how far this stock might go up, or how it’s doing, or about any news or information of any kind on this stock, the only question I care about in every trade every day is whether my stop should stay where it is, or be raised. Otherwise, I don’t think or even read about any of my current positions at all. This is a case where ignorance really is bliss, because irrelevant “information” is disinformation, and can only hurt my trading. Due to the “Confirmation Bias” discussed in a later chapter, the “information” that a trader tends to find will strongly tend to confirm his or her preconceptions, which is another way of saying close your mind to all of the information that does not conform to your preconceptions. I’ve already done all of the thinking that can help or hurt these trades, so I spend the rest of my time screening for other similar candidates.

e. Always enter your stop the instant that your buy order is filled. Make it good till canceled. NEVER wait to do this. It is inevitable that at least an occasional trade totally collapses, and with blinding speed. I’ve had a couple of stocks drop nearly 50% below my entry point—and well below my stop– within a day of entry, and one, within about 3 minutes. Setting a cost limit for each position will keep this from being catastrophic for your overall account. The stop loss order will keep it even smaller, in most cases. You will have some of these worst trades. The one time you go in big on a position and without a stop will be the one time that catastrophe strikes, one that you may never recover from. Failure to trade small (take small positions) and use a stop on all positions is the same as playing Russian roulette. Even if you’re playing with a gun with 100 chambers, one round has your name on it. You could be right 99% of the time on your trades and still get wiped out. If this happens on your first trade, you don’t get to benefit from being right 99% of the rest of the time, because you’re already out of the game. A no-stop strategy will eventually lead to ruin for essentially all traders. “This means that once you’re in a losing trade, you don’t have to do anything to keep on losing. You don’t even have to watch. You can just ignore the situation, and the market will take everything you own.”(6) A “mental” stop that has been calculated but not entered is not a stop. Don’t worry about givin