POSITION SIZING RULE: No matter which method you adopt make sure it incorporates your expected trade risk, “R”, based on your initial stop-loss. Determine your size with your R and total risk which I recommend to set at 1% of your total capital.
Gary Brinson and associates published a landmark paper in 1986, “Determinants of Portfolio Performance.” They found that 91.5% of the difference between one portfolio’s manager performance and another portfolio manager’s performance are best explained by asset allocation. Since then several other authors have revisited his conclusions.
Vanguard reviewed these studies, “The Asset Allocation Debate: Provocative Questions, Enduring Realities” and made the following conclusions:
- Broadly diversified portfolios tend to move in tandem with the market over time.
- Active management creates significant performance dispersion across portfolios.
- On average, active management reduces a portfolio’s returns and increase its volatility compared with a static index portfolio. However, active management creates opportunity for a portfolio to outperform the market.
- The variability of returns can be explained largely by asset allocation policy.
While these studies directly studied how much in cash, bonds, or stocks, it is still a “how much” question. As a small investor it directly applies to how much we put in different stocks and how much we purchase of each stock.
How important is “How much?” Dr. Tharp has a game that he plays at his lectures on position sizing. He has a bag of marbles that a trader randomly pulls from. They are labeled in “Rs” which represents the dollar risk per trade. If you are willing to accept $100 loss in a trade, that is your R, $100.
In Dr. Tharp’s bag there are seven marbles representing 1R losers, one marble is a 5R loser, and two are 10R winners. The overall game has a 0.8R expectancy even though 80% of the marbles represent losing trades. The audience has $100,000 in equity and they are free to risk whatever they feel is appropriate before the marble pull for a total of 40 trades. As you guessed, each participant had different outcomes even though they are playing with the same random marbles. The outcome ranged from bankrupt to over $1 million in 40 trades.
The difference in performance for these investors playing Tharp’s game is psychology and position sizing. If you buy too much stock for the size of your capital you run the risk of ruin. If you don’t buy enough stock, you fail to take advantage of the winners and your portfolio never grows. There are many different strategies for sizing your position. Here are a few direct from Dr. Tharp’s book.
MODEL 1: Buy one unit for a fixed amount of money
The simplest method. A unit can be 100 shares or one contract. Essentially you buy one unit per fixed amount of money, say $20,000. So, if you have $20,000 you buy 100 shares. The poor trader with this system, because he does not know better, starts buying one contract of everything that looks promising. At $40,000 he graduates to 2 contracts (or 200 shares) of everything. Irregardless of the stock price. 200 shares of Apple at $105 or 200 shares of PayPal at $38, they are all the same to this trader. Obviously this is the worst performing model.
MODEL 2: Divide your account into equal value units.
Divide your capital into equal units. Say you have $100,000 and you want to be in 10 different stocks. You divide $100,000 by 10 and have 10 equal units of $10,000. You then buy as many shares as you can for $10,000.
Some professional traders and mutual fund managers use this method to control their initial position and use it for rebalancing. Unfortunately, it violates the golden rule of trading – This strategy means that you are trimming your winners and adding to your losers.
MODEL 3: A percent risk model
Once you enter a trade you need to know at what point you call “uncle.” What is the most you are willing to let this stock or trade go against you? Your worst-case scenario, your “R” or risk.
Let’s use NFLX as an example. You have been watching the chart and you notice a double top at $102 (really $102.35 but we will round to $102 to make the example easier). Then on March 29 NFLX breaks out and closes above $102. The next day it opens above $102 so you pull the trigger and enter. You see a pivot point at $97 and this becomes your stop-loss point. So your risk, or “R,” is $5. How many shares should you buy?
Your account is $100,000 and you decide to limit your total risk to 2% of your account or $2,000. $2,000 divided by $5 equals 400 shares. Your total cost would be 400 shares times $102 or $40,800. So your total investment is nearly 41% of your account, but your total risk is only 2% of your capital.
Now we follow NFLX and we sell at the peak, boy we are good, at $112 (actually $111.85). We made $10/share ($112 – $102). Our risk, “R”, was $5/share. So we say that this trade was a 2R trade ($10/$5 = 2). And, with 400 shares we made $4,000! Not bad.
The NFLX example shows that 2% risk on a $100 stock limits you to around 2 stocks. So, you can either lower your risk, choose lower priced stocks or get more capital.
MODEL 4: The percent volatility method.
Volatility is the amount of movement over a set period of time. We usually use a day as the period of time. The volatility is the high of the day minus the low of the day. If NFLXs daily high is $100 and the low of the day is $97 then the volatility is $3. If you take a couple days, including any gaps, and average them together you have the “average true range” or ATR.
Back to NFLX. On the day we bought the stock at $102 the ATR(20) (a 20 day average ATR) is $3.83. We will round up to $4 for this example. Now we can buy 500 shares if we keep a 2% risk ($2,000/$4).
What makes this system particularly good is that as the volatility of the stock increases, you take smaller size. Less volatility means a larger size.
Kelly criterion money management
John Kelly worked for AT&T’s Bell Laboratory and developed the Kelly Criterion to assist AT&T with long distance telephone signal noise. It was published as “A New Interpretation Of Information Rate” in 1956. Very quickly the gambling community adapted it to maximize the size of their bankroll.
There are two basic components, the Win probability (W) and your Win/loss ratio (R). The equation is then: Kelly% = W – [(1-W) / R]
Here is how you can use it. You need your past performance for the last 50 or so trades.
- W = # winning trades / # losing trades.
- R = the average gain of the winning trades / average loss of the negative trades
The final Kelly% is your risk position.
I bring this method up as you may have heard about it. It is really meant for gaming like blackjack and I probably would not use it for stocks.
Systems used by two of the world’s greatest traders from Trade Your Way to Financial Freedom.
William O’Neil’s CANSLIM Method
O’Neil does not describe any method of how many shares to own, what he does recommend is limiting your number of stocks. He feels that even a multimillion-dollar portfolio should only own six or seven stocks. Under $100,000 limit to four or five stocks. Under $20,000 limit to two.
His only interest is owning a few of the very best businesses. He wants to own as much of these few businesses that he can.
Well, I suppose not many of us have unlimited bank accounts, so let’s stick to a system that limits our risk.
I hope I have convinced you that position sizing is one of the most important aspects of trading and accounts for over 90% of the variability between professional portfolio managers. A good sizing method and a good exit strategy practically guarantees success.
There are hundreds of sizing methods. A great book to read is by Dr. Tharp, the expert in position sizing, titled, “Van Tharp’s Definitive Guide to Position Sizing.” No matter which method you decide to use for your own personal trading make sure you incorporate the concept of risk per trade, R, and set a stop-loss. Also determine what percent of your total capital you are willing to risk on any one trade.
There is never a one-size-fits-all technique. In the example of NFLX there was a readily identifiable pivot point. However, it could be the 200 moving average, a double bottom or a Fibonacci line. If nothing is readily identifiable within a reasonable distance from the stock’s current price you could use volatility. You need to be adaptable, no matter how you choose your stop-loss, always know your R and what percentage of your total capital you are willing to risk to determine your size.
If you are a novice or very conservative trader, start with 0.5%. If you are a professional and have strong convictions on a stock you can risk up to 2%, but I recommend limiting your risk to 1%.
Remember, this is your initial position size and stop-loss. This does not address how to manage the trade or determine your exit(s).
Because this is such an important issue I will be writing future articles in much more detail explaining how different methods suit different situations.
NOTE: As you know I am not a strong advocate of indicators. But, you do need to determine your stop-loss and to do this you need something to go by. So while I am more visual in determining when to enter a trade, I do use some simple indicators to help determine a stop-loss and manage the trade.