Stock Learning

Part 1 discussed three equity models: total equity, core equity, and reduced total equity. It also discussed four money management models: 1) the one unit per so much equity; 2) the leverage model; 3) the percent margin model; and 4) the percent volatility model. The four models can be combined with each of the equity models to produce 12 different money management models - even more if you combine them or add creative money management.


Model 5 - Percent Risk

When you enter a position, it is essential to know that point at which you would get out in order to preserve your capital. This is your “risk.” It’s your worst case loss - except for slippage and a runaway market going against you. One of the most common money management systems involves controlling position size as a function of this risk. Let’s look at an example of how this money management model works. Suppose you want to buy gold at $830 per ounce. Your system suggests that if gold drops as low as $820, you need to get out. Thus, your worst case risk per gold contract is 10 points times $100/point or $1.000. You have a $50,000 account. You want to limit your total risk on your gold position to 2.5% of that equity or $1,250. If you divide your $1,000 risk per contract into your total allowable risk of $1,250, you get 1.25 contracts. Thus, your money management using model 5 will only allow you to purchase one contract.
Suppose that you get a signal to sell short corn the same day. Gold is still at $830 an ounce, so your account with the open position is still worth $50.000. You still have $1,250 in allowable risk for your corn position based upon the total equity model. Lets say that corn is at $4.93, and you decide that your maximum acceptable risk would be to allow corn to move against you by 5 cents to $4.98. Your 5 cents of allowable risk (times 5,000 bushels per contract) translates into a risk of $250 per contract. If you divide $250 into $1,250, you get 5 contracts. Thus, you can sell short 5 corn contracts within your money management paradigm. In these examples, we’ve used a total equity model to calculate our risk, where total equity refers to the cash value of the account plus the value of all open positions. In contrast, lets see what would happen if we used a core equity calculation of risk. In the core equity model, the risk involved in open positions is subtracted from the cash value when those positions are opened and only the remaining cash value is used in subsequent calculations. First, we purchased a gold contract and our total risk exposure in that contract was $1,000. In the core equity model, our new core equity is $1,000 less. Thus, we only have $49,000 left on which to base the risk for our next position in corn. Since our money management allows us to risk 2.5% of this core equity, we can risk $1225.

We now want to sell short corn with a risk of $250 per contract. If you divide $250 into $1,225 you get 4.9 contracts. Thus, the core equity model would only allow you to sell short 4 corn contracts. Notice
that to be conservative and not exceed our parameters, we always round down to the nearest whole unit. Lets say that your next purchase of corn isn’t the same day. You get your signal six weeks into the future. You still have an open position in gold, but now gold is $940 per ounce. Thus, your open position is worth $11,000. As a result, your total equity is now $50,000, plus the value of the open position, or $61,000. If you are using the total equity model, you can now risk 2.5% of $61,000. Therefore, you could now risk $1525. If the corn signal occurred with $250 risk per contract, your money management would now permit you to sell short 6.1 ($1525 divided by $250) contracts. In contrast, the core equity model would still be based upon $49,000 and would only allow you to sell short the same 4 contracts of corn. Obviously, of the three equity models, the core equity model is the most conservative. Reduced total equity ranks in the middle, and the total equity model is the most risky model.

How does the percent risk money management compare with the percent volatility money management discussed in the last issue? Table 4 shows the 55/21-day breakout system (used as an illustration
in Tables 2 and 3) with a money management algorithm based upon risk as a percentage of equity. The starting equity is again $1,000,000.

Table 4: 22:21 Breakout System with Risk Money Management

If you compare Table 4 with Table 3 from Money Management Part 1, you’ll notice the striking difference in the percentages at which the system breaks down. These differences are the result of the size of the number (i.e., the current 21-day extreme against you versus the 20-day volatility) that you must take into consideration before using the equity percentages to size positions. Thus, a 5% risk based upon a stop of the 21 day extreme appears to be equivalent to about 1% of equity with the 20 day average true range. These numbers, upon which the percentages are based, are critical. They must be considered before you determine the percentages you plan to use to size your positions. Notice that the best reward-to-risk ratio occurs at about 25%, but you would have to tolerate an 84% draw-down in order to achieve it. In
addition, margin calls (which are set at current rates and not historically accurate) start entering the picture at 10% risk. If you traded this system with $1,000,000 and used a 1% risk, your bet sizes would be equivalent to trading the $100,000 account with 10% risk. Thus, Table 4 suggests that you probably should not trade this system unless you had at least $100,000 and then you
probably should not risk more than about 0.5% per trade. And at 0.5%, your returns with the system would be very poor. Essentially, you should now understand why you need at least a million dollars to trade
this system.
Just how much risk should you accept per position with risk money management? Your overall risk using risk money management depends upon the size of the stops you’ve set to preserve your capital and the expectancy of the system you are trading. For example, most long-term trend followers use trailing stops that are fairly large, several times the average daily range of prices. In addition, most trend followers are usually using a model that makes money 40-50% of the time and has a reward-to-risk ratio of 2.0 to 2.5. If your system does not fall into these ranges, then you need to determine your own money management percentages. With the above criteria (and precautions) in mind, if you are trading other people’s money, you probably should risk less than 1% per position. If you are trading your own money, your risk depends upon your own comfort level. Anything under 3% is probably fine. If you are risking over 3%, you are a “gunslinger” and had better understand the risk you are taking for the reward you seek.
If you trade a system that sets very small stops, then you need to adopt much smaller risk levels. For example, if your stops are less than the daily range of prices, then you probably need guidelines that
are about half (or less) of what we present here. On the other hand, if you have high expectancies in your system (your reliability is above 50% and your reward-to-risk ratio is 3 or better), then you can probably risk a higher percentage of your equity fairly safely. People who use very tight stops might want to consider using a volatility model to size their positions.
Most equity traders don’t consider this sort of model at all. Instead, they tend to think more in terms of the equal-units model. But let’s look at how “risk money management” would work with stocks. Let’s say you want to purchase IBM and you have a $50,000 account. IBM’s price is about $126 per share. You decide that you would get out of this position at $122, or a drop of $4 per share. Your money management routine tells you to limit your risk to 2.5% or $1250. When you divide 4 into 1250 you come up with 312.5 shares. If you bought 312 shares at $126, it would cost you $39,312 - over half of the value in your account. You could only do that two times without exceeding the marginable value of your account. This gives you a better notion of what a 2.5% risk really means. In fact, if your stop was only a $1 drop to $125, you could purchase 1250 shares based upon the model. But that 1250 shares would cost you $157,500 - which you couldn’t do even by fully margining your account. Nevertheless, you are still limiting your risk to 2.5%. The risk calculations, of course, were all based upon the starting risk - the difference between your purchase price and your initial stop loss.


Model 6 - Periodic Money Management Adjustments

Consider monitoring your money management on a periodic basis - weekly, daily, or even hourly - to maintain a fairly constant exposure. Think about the potential here. You could monitor each position and make sure that your exposure was always 1% or less. This means that, except in runaway markets, your biggest risk would always be about 1%. Your exposure could be monitored using any of the money management models given or any of the equity models suggested. However, I would suggest that you consider monitoring both ongoing risk and ongoing volatility with a total equity calculation. Here’s how daily monitoring for risk and volatility might work. Lets suppose you have a $200,000 account and you have open positions in gold and corn. Your money management says you will keep your initial risk to 2% of equity and your ongoing risk at 3% of equity. You’ve purchased four long gold contracts at $900 per ounce with a stop at $890, so you now have open risk of $1,000 (i.e., 10 points times $100 per point) per contract, or $4.000. The next day at the close you monitor your open risk. Lets say gold has jumped to $940 overnight. Your gold stop is now $910. The $40 increase in gold has increased your equity by $16,000 (i.e., 4
contracts times 40 points times $lOO/point). Thus, your total equity is worth $216,000. Your open risk for gold is now at $30 (i.e., $940 less $910) per contract. The total value of that open risk is $3000 (i.e., 30 times $100 per point) per contract or $12,000.
You have decided to monitor your open risk on a dailv basis and keen it at 3% of total equity. Doing so still allows you to follow your trading model. More importantly, it reduces the chances of any large
declines in equity occurring in a short period of time. Since 3% of $216,000 is $6,480, you can now only afford to keep two gold contracts. You must sell off the remaining two contracts. Some of you might say, “why not raise your stop so that you could keep the four gold contracts?” Remember, money management is a separate part of your system that tells you how much. If you altered your stop, you wouldn’t be following your trading system which now says that your stop should be at $910 - your exit and your money management would start to merge. By selling two contracts, you are simply reducing your risk in order to keep your total risk within acceptable limits on a daily basis according to your money management guidelines. You still have the opportunity to profit if gold keeps moving in your favor and you won’t be giving back as much of your profits should gold suddenly decline. Thus, you are making a money management decision to maintain a constant risk in your portfolio.
Let’s see how the same adjustments might occur with volatility. Suppose you have a $200,000 account and you decide to buy corn at $5.00. Your model says that you will buy enough corn so that the daily volatility of corn was only 1% of your total equity. In addition, you will never allow the daily volatility to go beyond 2% and you elect to monitor daily volatility each Monday. Assume that the daily volatility was 8 cents when you purchased it. This translates into a price range of $400 per day (i.e., 5,000 bushels x 8 cents/bushel = $400). You decided not to allow volatility to exceed 1% of your $200,000 equity or $2,000 when you purchased the corn, so you bought five contracts.
Suppose corn jumps to $6.00 over the next month so that your five corn contracts have given you a profit of $25,000. The daily volatility of corn is now 20 cents. Since your total equity is now $225,000, you
can now allow your daily equity to fluctuate by 2% of that amount or $4500. However, corn volatility is now $1000 per contract. You have five contracts, giving you a total volatility of $5,000. As a result, you must sell one corn contract according to the criteria of your periodic volatility money management model.
Generally, when something begins to increase in price dramatically the volatility will also go up dramatically. If you are in such a move, you might find that you have a $100,000 starting account that’s now
worth $500,000. In addition, because of the large increase in the daily price volatility, you might find that your account changes value by as much as $100,000 each day. By keeping a volatility adjustment as
part of your money management, you protect your open profits and prevent such large daily fluctuations in your account. You read examples of periodic monitoring of your money management for the risk and volatility models. However, you can do periodic monitoring with all of the models mentioned. You can even do a combination of them, such as monitoring risk and volatility simultaneously. Are you beginning to see the possibilities?

Model 7 -Group Control

One of the most important factors in risk control is having a diversified portfolio. Trading a number of items generally spreads your risk around, provided that price changes in those items have a low-correlation.
Here’s how Group Control works. Suppose you are trading a system that makes money in 5 of 12 trades. The average winning trade is about 2.5 times the size of the average losing trade. In addition, the system only generates about one trade per month per investment vehicle. If you only traded one instrument you would have about one trade each month. This means your chances of having a winning month are only about 41.7%. You could easily have six months of losses and become discouraged. Suppose that you trade 10 different instruments that are all independent of each other. Each one of them, let’s say, is likely to generate a trade each month. Table 5 shows 1) the number of losing trades out of 10 you might have; 2) the chances of that happening; and 3) the amount of money you’d make or lose on that combination assuming an equal unit risk on each trade and a 2.5-to-l reward-to-risk ratio.
Notice that you would need to have less than three winning trades out of ten in order not to make money. The chances of that occurring on a given month (in which you have 10 trades) is the sum of the first
three probabilities or 14.2%. Thus, with 10 independent markets, you only have a 14% chance of having a losing month.

Table 5: Possible Results with 10 Independent Units

When you try to put this plan into effect, however, you run into the difficulty that most trades are not independent. Stocks tend to go up and down together. During bull markets, there are tendencies for certain groups of stocks to move together. For example, much of the stock market move in 1999=2000 was due to technology stocks. When the
move was over, technology stocks tended to fall as a group - and 2000-2001, most of them did.
Commodities also tend to have groupings that are highly correlated. Grains, metals, meats, stock indices, currencies, energies, etc. might each tend to move as a group in the same direction at the same time.
Thus, your goal through money management is to minimize the number of highly correlated positions in your portfolio at any given time. You could do this by preselecting a limited number of vehicles in which to invest or trade. This is the portfolio selection part of system design.
However, you can also accomplish this diversification by having a money management algorithm limiting your total group exposure by using one of the methods presented so far. For example, you could limit the amount of leverage in any one group. You also could limit the amount of risk, volatility, leverage, margin, or total number of units of exposure that you have in any one group. This has the advantage of
limiting your group exposure, while avoiding the possibility of missing a good opportunity because it is not part of the portfolio which you have preselected to trade.
Suppose your overall money management algorithm is to limit the new risk on any given position to 1% of equity. Your model calls for you to trade any liquid commodity that tends to fit your trading model.
When you do that, however, you might find yourself with a portfolio of US bonds, 10 year notes, T-Bills, Euros, Muni-bonds, German Bonds, etc. That wouldn’t be prudent, because your entire portfolio would be controlled by interest rate fluctuations. As a result, you decide to limit your total group risk to 3%. Based upon your initial risk allocation, the most you could have is three 1% positions in any one
commodity grouping. Note that your group money management model could be based upon any of the first five models presented - the one unit per so much equity model, the leverage model, the margin model, the
volatility model, or the risk model.


Model 8 - Portfolio Heat

It’s also important to limit the total risk to which your portfolio is exposed. This value has been called portfolio heat by Ed Seykota and Dave Druz in Determining Optimal Risk. Technical Analysis of Stocks and Commodities. Most great traders would argue that 20- 25% is probably a maximum level for your portfolio heat. However, portfolio heat should also depend upon how good your system is. For example, a 60% system with average gains that are 4 times the size of average losses could have a much greater portfolio heat than a 50% system with a 2-to-1 gain-to-loss ratio.
A good rule of thumb for determining your portfolio heat is to calculate the Kelly criterion for your system. The Kelly criterion gives you a good approximation for the maximum risk possible for your system. 80% of that value is probably a good number to pick for your total portfolio risk. However, if 80% of the Kelly criterion for your system is still above 25%. you could be flirting with danger. Once you have a number in mind for your portfolio heat, work backwards to determine the individual risk on any given position. How many positions are you likely to have on at any given time? Take your maximum number of positions and divide that into the number you’ve just calculated for your portfolio. That’s probably a good estimate for the maximum amount of risk you should assume for a single position.
However, these guidelines also make the assumption that you are going for maximum gains in your portfolio. Portfolio heat was a term coined for the total risk of your portfolio. However, you could apply any of the first five models, or a combination of them, to your total portfolio. Notice how money management is getting more complex and more sophisticated as we add more models.


Model 9 - Long versus Short Positions

Several famous traders have distinguished between long and short positions in considering group risk and portfolio heat. They believe that they somewhat counteract each other, so that one long position and one short position - each at your desired market money management level - would just need to be counted as one unit. In other words, a 1% risk in a long corn position and a 1% risk in a short bond position might be grouped together as one 1% unit of risk. This puts an interesting twist to many of the money management models already presented.
Equating different long and short positions, of course, can only be used with those models which equate your exposure. Thus, it would not be applied to Model 1, but you could apply it with Models 2 through 5.