How to Develop Winning Trading Systems

That Fit You Workshop

Dr. Van K. Tharp, PhD &

RJ Hixson

Recorded Video Handout

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Position Sizing Strategies

Objectives

1. To gain an overview of basic position sizing strategies, including CPR for traders.

2. To learn some general guidelines for position sizing strategies depending on your objectives.

3. To learn several strategies used by great traders to get maximum returns.

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Position sizing strategies are the part of your trading system that answers the question “how much” throughout the course of a trade.

The purpose of position sizing strategies is to help you achieve your objectives.

Effective position sizing strategies require two inputs - 1. Trading Objectives – Are they well defined? Have both components?

• Drawdown – avoid (plus your perceived probability) • Returns – achieve (plus your perceived probability)

Possible Trading Objectives

• Make as much money as possible without worrying about drawdowns or even ruin

• Maximize the probability of achieving your returns • Maximize the probability of meeting returns objective while minimizing

probability of hitting your max drawdown (or some level of drawdown) • Minimize the probability of ruin or even hitting a drawdown of X% • Asset allocation for systems/asset classes

By assigning figures to these variations, you can see that a HUGE number of objectives (perhaps infinite) is possible. Each set of objectives would have a position sizing algorithm that would maximize the likelihood of achieving those objectives.

2. System Performance

SQN is easiest single measure for development of your position sizing strategy Higher SQN scoring trading systems makes it easier for position sizing

strategies to achieve your objectives Lower SQN scoring systems make it more challenging (for the position

sizing strategy and for trading) to achieve your objectives

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Understanding CPR for Traders

This figure shows the relationship between the Risk (R) per unit, the Position Size (P) for the trade, and the total equity or Cash (C) at risk in a trade. With any of the two variables the third can be calculated by following the “Y”.

P = C / R Or P = RT / RU

We’ve shown it this way so you don’t have to memorize the formulas. Just remember these simple mnemonics that you can use to remember the Y diagram:

“I learned CPR at the Y.” or “Cash, Position Size, Risk.”

Cash (the equity you are willing to lose on the trade in $ or RT)

Risk (difference between entry price and initial stop in $ per unit or RU)

Position Size (# of units) =

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Some Position Sizing Strategy Basics Martingale vs. Anti-Martingale Systems When applying % of equity strategies - how do you define equity? There are three ways:

Total Equity: Equity is cash plus equity amounts in open positions

Reduced Total Equity: Equity is cash plus equity in open positions that stops protect.

Core Equity: Equity is available cash only—no portion of open positions are included.

Some Very General Equity Risk Percentage Guidelines

0.1% to 1% per trade when managing other people's money 0.8% to 2% per trade when trading your own money 3% or greater (according to Ed Seykota) is being a gunslinger

Additional Position Sizing Strategies

Percentage of Volatility Percentage of Margin Percentage of Leverage One Unit per So Much Equity Scaling In Scaling Out Group Risk Portfolio Heat—Total Portfolio Heat should not exceed 20% to 25%

Going for the Moon Strategies

Recommended - Market’s Money/Creative Position Sizing Strategies Not Recommended - Kelly Criterion Percentage, Optimal f: (What fraction of your maximum loss should you bet?), Fixed Ratio Trading Cop

yrigh

t (c) 2

021 V

an Th

arp In

stitut

e

Some Position Sizing General Examples For Different Objectives

Objective 1: Make as much money as possible without worrying about drawdowns or ruin.

Method 1: Risk the % Risk that Gives You the Highest Median Gain

System Quality Number score above 5.0 – could do as much as 4%, but be careful of multiple correlated positions.

System Quality Number score above 3.0 – could do as much as 2.5%, but again be careful of multiple correlated positions.

System Quality Number score above 1.75 – could do as much as 1.5%, but be careful of multiple correlated positions.

System Quality Number score below 1.75 – with this objective, trading this kind of a system is probably a disaster

Method 2: Bill Eckhart Pyramiding Start at ½-1% of equity and as the price goes up by a certain amount – which you define, scale in 3-4 times. Limit, however, your total exposure to the maximum amounts suggested in Method 1.

Objective 2: Maximize probability of meeting goals.

Use Percent Risk that Gives Highest Probability of Meeting Your Objective

System Quality Number score above 5.0 – could do as much as 2.5%, but be careful of multiple correlated positions.

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

System Quality Number score above 3.0 – could do as much as 2.0%, but

again be careful of multiple correlated positions.

System Quality Number above score 1.75 – could do as much as 1.25%, but be careful of multiple correlated positions.

System Quality Number below score 1.75 – with this objective, trading

this kind of a system is probably a disaster Objective 3: Maximize the probability of meeting your goal, while minimizing any loss to your initial equity.

Method 1: Market’s Money

On starting equity use % risk that has < 0% chance of ruin. But on profits, use risk percent that is optimal for achieving goal or use median gain %.

Here are some examples for your starting risk (i.e., on your core equity before market’s money):

System Quality Number score above 5.0 – could do as much as 1%, but be careful of multiple correlated positions.

System Quality Number score above 3.0 – could do as much as 0.8%,

but again be careful of multiple correlated positions.

System Quality Number score above 1.75 – could do as much as 0.5%, but be careful of multiple correlated positions.

System Quality Number score below 1.75 – could probably trade this

with 0.5%, but would not recommend using much more than 0.75% with Market’s Money.

For the Market’s Money portion of your position, use the examples listed above for Objective 2.

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Method 2: Use a Market’s Money strategy, but convert the market’s money to equity every month.

This is more conservative than Method 1. Method 3: Scaling-Out

Scale-Out to maintain constant open risk Scale-Out to maintain constant volatility

This method could require monitoring your position sizing on a periodic basis—weekly, daily, or even hourly—to maintain a fairly constant exposure. What potential risk are you exposed to?

• Here you need to calculate the difference between the current value

and the stops of each position you have. This is called the open risk of your portfolio. What if you controlled the total open risk or limited the open risk of each open position by scaling-out of positions to limit your open risk? Think about the potential here. You could monitor each position and make sure that your exposure was always 2% or less per position. This means that, except in runaway markets, your biggest risk would always be about 2% per position. You must have a System Quality Number score of at least 2 to do this.

• In addition, think about the potential volatility of each of your positions. What has the volatility been in the positions you hold over the last few days? Is this volatility going up? What if you limited it by scaling-out of positions when a certain level of volatility has been passed?

• Consider monitoring both ongoing risk and ongoing volatility with a

total equity calculation.

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Objective 4: Minimize probability of ruin or even a drawdown of X%.

Method 1: Percent Equity Risked:

System Quality Number score above 5.0 – could do as much as 1%, but be careful of multiple correlated positions.

System Quality Number score above 3.0 – could do as much as 0.8%,

but again be careful of multiple correlated positions.

System Quality Number score above 1.75 – could do as much as 0.5%, but be careful of multiple correlated positions.

System Quality Number score below 1.75 – could probably trade this

with 0.5%, but would not recommend using much more than 0.75% with Market’s Money.

Method 2: Calculate your average maximum R-drawdown and use that to calculate position size.

Let’s say your objective is to avoid a 25% drawdown at all costs. And here, we are talking about a drawdown at any time in the curve of 25%—not just a drawdown of 25% from the starting equity.

The way we’ll do that is to calculate the maximum drawdowns of our system in terms of R. Let’s simulate a system making 100 trades 10,000 times to see what is possible in terms of drawdowns expressed in terms of R.

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

The median maximum drawdown was 38R. But in 1,000 of the simulations (i.e., the 10% level), we had a maximum drawdown of 62R. The table below shows other probabilities from the simulator from this particular system.

Probability of Maximum R-Drawdowns in Our System Maximum Drawdown Probability DD

-12.1R 100% -29.1R 76.4% -38R 50% -48R 25% -60R 10% -71R 5% -93R 1%

Your system, depending upon its R-multiple distribution, will have different probabilities. Cop

yrigh

t (c) 2

021 V

an Th

arp In

stitut

e

Let’s use the 10% level of -62R to do our calculations. Thus, we can feel with certainty that we have less than a 10% chance of reaching these levels. If we divide 25% by 60R, we get a risk level of 0.4%. This is pretty similar to our estimate of our initial risk size. However, if we want to only have a 10% chance of a 25% peak-to-trough drawdown in our equity curve, then we must never risk more than 0.4% with this system. And if we wanted to guarantee a 1% or less chance of such a drawdown, we probably shouldn’t risk more than 0.2% (i.e., 25% drawdown/93R = 0.00269).

From The Definitive Guide to Position Sizing Strategies

Expected Losing Streaks in 100 Trades (or Losing Streaks as a Function of Winning Percentage of Trades)

System Winning Trade

Percentage

100% Probability for a Losing Streak

This Long

Average Losing Streak Length

10% Probability for a Losing Streak

This Long

1% Probability for a Losing Streak This

Long

Maximum Losing Streak

Length

80% 2 3 4 5 to 6 7 75% 3 3 5 6 to 7 9 70% 3 3 5 to 6 7 to 8 10 65% 3 4 6 to 7 8 to 9 13 60% 4 5 7 9 to 10 14 55% 4 5 8 10 to 11 16 50% 5 6 9 12 19 45% 6 7 10 13 to 14 22 40% 7 8 11 to 12 15 to 16 25 35% 8 9 13 to 14 18 to 19 34 30% 9 11 15 to 16 22 38 25% 10 13 18 to 19 25 to 26 41 20% 12 15 22 to 23 32 51

Method 2 Summary Steps-

1. Determine the worst-case peak-to-trough equity drawdown you’d like to avoid.

2. Simulate your system and determine the probability of various R-level

maximum drawdowns. 3. Determine the maximum drawdown in terms of R at the probability level

you are willing to accept.

4. Divide this level into the value you selected at step one and the result should be the position sizing level as a percent risk that you should use.

Cop

yrigh

t (c) 2

021 V

an Th

arp In

stitut

e

Objective 5: Belief: Asset Allocation and Position Sizing are the same—they both are designed to answer the question of “how much.” Method 1: Asset Allocation for Classes of Investments. Suppose you want to always hold some positions in the strongest markets or market sectors. You could do this by buying ETFs and staying in those positions until they show themselves to be weaker than the S&P 500 or until they are out of the top relative strength positions. One good way to pursue this strategy is to allocate a percentage of your total capital to this strategy (i.e., such as 20%). This way you could invest in the five strongest sectors. Method 2: Asset Allocation between Systems. Tom Basso and Jack Schwager have shown that monthly or quarterly rebalancing works to improve the performance of non-correlated traders, then it should also work with non-correlated systems. If you have 3 different systems, allocate money based on the following: The System Quality Number score of each system, with the best system getting the most money.

o 5.0 SQN = 60% o 3.0 SQN = 30% o 1.75 SQN = 10%

If the systems have similar SQN scores, allocate equally between them.

Each month (or quarter), rebalance your money between the systems based upon the initial allocation.

In Every Case of Calculating Your Position Sizing Risk Amount -

Remember that you also need to manage your total risk at the portfolio

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

level. Numerous open positions in the market probably constitute multiple correlated positions. Multiple correlated positions can act like one big position when exogenous events happen – outside events which affect all positions in the market simultaneously. We refer to total portfolio risk exposure as Portfolio Heat. Rough Guideline for Portfolio Heat Amounts

SQN Score above 5 up to 20% SQN Score 3 - 5 up to 15% SQN Score 1.75 - 3 up to 12% SQN Score 1.75 up to 6% SQN Score 1.3 - 1.75 up to 3%

Example – You trade a system with a SQN score of 3 and based on your objectives, you intend to risk 1% of your equity on each position. You also know that you can have up to 24 positions open at any one time. Using the portfolio heat guidelines above you realize you would not want to have more than 12% open risk in the market at any one time. Therefore, you would then adjust your risk amount per position from a max of 1% down to a max of .5%.

12% portfolio heat / 24 open positions = max .5% risk per position

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Strategies of the Masters Ed Seykota’s Market’s Money Strategy When you size to make the optimal rates of return, you risk big drawdowns, which is usually not acceptable. However, if you only position size for optimal rates of returns on the equity you’ve earned from the market (i.e., the market’s money), then it’s possible to still have huge returns without risking a large drawdown to your starting equity. For example, you might determine that the optimal risk size for returns is about 5%, but that anything over 1% risks unacceptable drawdown. Thus, you might risk 1% of your starting equity and 5% of the market’s money. For example, if you started with $100,000, then the starting position size would be based on 1% of $100,000 or $1000. However, two months later your total equity might be $120,000. Now you can risk 1% of $100,000 (i.e., $1000) plus 5% of $20,000 (i.e., the markets money), which is also $1000. Thus, your new position size would be $2000. Your equity has only gone up by 20%, but your position size has doubled. It’s possible to make huge rates of returns using a market’s money philosophy. The only question you need to answer is “WHEN DOES THE MARKET’S MONEY BECOME MY MONEY?”

• At the end of the year when you get taxed on it. • At the end of the month when you must report to your investors. • At any other time period. • After your account has increased by a desired percentage. • After your account has increased by a desired amount of money.

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Tom Basso - Scaling Out Risk Control 1. New Position Risk

First, use a decision making model to decide the price at which you will enter the market. Make sure you have a plan to get out of the trade if you are wrong. This plan should include some sort of stop loss point to protect you from catastrophic loss. Convert the distance from the stop into dollars per contract.

Example: Buy gold at $400, with a stop price of $390, and the objective being as high as it wants to go. Let's trail it with a 10-day moving average of the closing price and use an account size of $200,000.

$RISK = (400 - 390) × $100/point = $1,000 risk per contract

Next we decide how many contracts to buy. I trade for clients and I prefer to limit my risk to 1% of equity. Lower risk levels will reduce your exposure even more. High risk requires more stomach lining and psychological control.

Allowable risk $ = 1% of $200,000 = $2,000 per position Contracts = $2,000/$1,000 risk per contract = 2 contracts

2. Ongoing Risk Exposure in an Existing Trade

Some investors/traders limit the risk of a new trade, but forget the psychological impact of higher risk associated with an existing trade. To control existing risk, I limit my risk in existing positions to 2.5% of equity. Continuing with the same example: Gold jumps to $450 overnight and our stop moves to $405. Our new equity is now $210,000. Where is our current risk and how do we maintain risk control?

Risk per contract = ($450 - $405) × 100 per point = $4500. Since we have two contracts we have $9,000 risk.

2.5% of $210,000 = $5,250 risk that we can allow. $5,250/$4,500 risk per contract (see above) = 1.167 contracts. Round this down to one contract. Therefore, it's necessary to sell one contract and keep the other one. Thus, every day the risk of your position is within a fixed range. Thus you are focused on the process of good trading, letting profits run and keeping risk to acceptable levels.

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Volatility Control Volatility is the dollar value of the true range of price movement for the day. The higher the amount of volatility, the more it will tend to distract you from your purpose of concentrating on good trading. Thus, I like to limit the amount of movement in any position to no more than 1% of equity on new and existing positions. This seems comfortable to my clients as well as me. More is a wilder ride, and less volatility is even more stable. Example: Gold is at $400 and the previous few sessions have averaged $3.00 range for an average day. Our equity is still $200,000. How many contracts would volatility allow us to do?

$3.00 × $100 per point = $300/contract

1% allowable volatility = 1% of $200,000 = $2,000 per position

Contracts = $2,000/$300 = 6.67 contracts. I'd round this down to 6 contracts.

Risk and Volatility Control Limit your position to the smaller of either risk or volatility. In the two examples shown, risk control yielded 2 contracts and volatility yielded 6, so I'd open the trade with 2 contracts.

Volatility in an Existing Trade I limit volatility in an existing trade to 1%, exactly like I calculate it for a new position. Then I just peel off the number of contracts necessary to get me back to an acceptable level of 1% equity. Cop

yrigh

t (c) 2

021 V

an Th

arp In

stitut

e

William Eckhardt’s Scaling In with Limited Risk For example, suppose you have a $100,000 account and you want to make your money grow as rapidly as possible. You are using a 3 times volatility stop as I did in the random entry trading system (reference Course Update #23a). You’ve also decided that your system is optimal risking 24% of equity at a time, using a reduced total equity model. You plan to have as many as six open positions at one time, so you are willing to risk up to 4% per position—but not all at once. You’ll build up to a position as big as 4% as your profits increase. Your initial risk will only be 2%. Let’s see how such a position sizing strategy system might work. You buy corn at $3.025. The ten day average true range (which we’ll call “V”) is 3.5 cents. Therefore, a 3 times volatility stop is 10.5 cents (i.e., at $2.92), which amounts to a total risk of $525. You can risk 2% of your $100,000, which amounts to 3 contracts (rounded down to the nearest contract). Your pyramiding scheme is to add one contract every time your profit increases by one daily volatility or V (which is currently 3.5 cents). When this occurs, (i.e., corn moves to $3.06) you risk another 2% with a 3 times V stop at $2.955. However, your stop on the original position moves up by 3.5 cents to $2.955. Thus, you now have six contracts all with stops at $2.955. However, notice that your total exposure of your original equity is now only 3% (actually less due to rounding) because you raised your initial stop. Let’s say that your daily volatility now increases to 4 cents. Thus, a new stop would now be 12 cents or $600. Corn moves up to $3.10, so you can now risk another 2%. (Actually, you could have done so at $3.095—when the price had increase by the old V-value of 3.5 cents.) Your reduced total equity is now $97,000 and 2% of that is $1,940. As a result, you can still purchase 3 contracts at $3.10—with a stop at $2.98. You also get to raise your stop on both of your other units by their respective V-values. Therefore you now have six contracts with stops at $2.99 and three contracts with a stop at $2.98. You might be saying, “How can you do that? Your risk is over the 3% limit with the reduced total equity model.” No, it isn’t because you raised your other stops enough so that your exposure is still about 3% of your reduced total equity.

Contracts Current Remaining Risk in Total Risk toStop Original Equity Original Equity

3 at $3.025 $2.99 3.5 cents 10.5 cents = $5253 at $3.06 $2.99 7 cents 21 cents = $1,0503 at $3.10 $2.98 12 cents 36 cents = $1,800

Let’s say that volatility stays at 4 cents and corn now goes to $3.14. It’s time to risk another 2%. Your reduced total equity is now $96,625. You can risk 2% of that or $1932.50. Your 12 cent stop is a $600 risk, so you can again purchase another 3 contracts. You must also raise your stops on the existing contracts. The stop on the first six contracts rises to $3.025 (i.e., it was raised 3.5 cents, the original V). The stop on the last three contracts rises to $3.02.

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Consider where you are with respect to the reduced total equity model in terms of risk. You now have risked 2% four times, but have you exceeded your 4% limit?

Contracts Current Remaining Risk in Total Risk toStop Original Equity Original Equity

3 at $3.025 $3.03 03 at $3.06 $3.03 3.5 cents 10.5 cents = $5503 at $3.10 $3.02 8 cents 24 cents = $1,2003 at $3.14 $3.02 12 cents 36 cents = $1,800

The total risk to your original equity is now only $3,550 or 3.55%—still under our 4% limit. So let’s say corn starts to really get volatile now and V goes to 6 cents. And you get a chance to buy more corn as it goes up to $3.20 (actually you could buy at $3.18, when it increased by the last value of V). But we’ll say that you buy at $3.20. Your total reduced equity is now $96,450 and 2% of that is $1,929. Your new stop, at 3 V, is now 18 cents or $900. Thus, you can now only purchase two contracts, but you also get to raise your other stops. Let’s say that we make a decision to leave the break-even stop alone, giving it plenty of room to move. However, you can now move the stop on the second 3 contracts purchased to break-even, move the stop on the contracts purchased at $3.10 to $3.06, and move the stop on the contracts purchased at $3.14 to $3.06. Thus, the current risk picture is shown below. Notice that by the reduced total equity model, your risk has changed very little. The risk to your original equity is now $3,600 or 3.6%.

Contracts Current Remaining Risk in Total Risk toStop Original Equity Original Equity

3 at $3.025 $3.03 0 03 at $3.06 $3.06 0 03 at $3.10 $3.06 4 cents 12 cents = $6003 at $3.14 $3.06 8 cents 24 cents = $1,2002 at $3.20 $3.02 18 cents 36 cents = $1,800

Let’s now say that corn goes as high as $3.50. Your exit gets you out at $3.40. Your profit is as follows.

Contracts Profit on Contract Profit in Dollars 3 at $3.025 37.5 cents $5,600 3 at $3.06 34 cents $5,100 3 at $3.10 30 cents $4,500 3 at $3.14 26 cents $3,900 2 at $3.20 20 cents $3,000

Total Profit $22,100 Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Your total profit was $22,100 and your maximum risk to your core equity was $3,600. Does that seem like a reasonable rate of return? Some of you might be saying “... but you ended up with 14 contracts! It might have been disastrous if you’d had some limit moves against you.” That’s true, but my point was to show you creative a position sizing strategy. In addition, there are other ways to protect against such limit moves (i.e., options) that make the risk well worthwhile. In fact, it is quite ironic that at the time I developed this example I was looking at a real corn trade. Corn actually went above $4.80 in that particular move. Had you really pressed this particular trade (as this position sizing strategy will allow) that trade would have been a trade of a lifetime. I haven’t figured it out, but you probably could have made as much as a million dollars without risking over $3,600 in your core equity. However, your risk to your total equity would have been considerable before the trade was over. There are any number of variables that you can vary in creative money management—your initial stop, your maximum risk per commodity, moving your stops in your favor, your equity model, your money management model, etc. For example, you could even use the idea of increasing your “reduced total equity” by raising your stop to justify opening up positions in other commodities. This could really help the small trader who does not have a large enough account to trade using most of these models. Contest Winning Strategies Using “Go for It Strategies” that produce the highest returns if you are lucky and survive the drawdowns: For example, one trade supposedly turned $10,000 into $2 million (lost some of it down to $1.1 million at the end of the contest). This trader was trading bonds with a volatility breakout system using optimal f.

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Kelly Criterion Once you have a trading system and have tested it out, you need to calculate three primary statistics to determine the amount of risk that will produce the maximum rate of return for you. The maximum rate of return will also tend to produce the largest drawdowns. These statistics include the reliability of the system, the size of the average gain, and the size of the average loss. Gambling has worked out a formula that you can use to determine your maximum risk size as a percentage of equity.

Kelly % = A – [(1 – A)/B]

Where A is the % of winning trades in decimal form (reliability of system), And B is the average profitable trade in $ divided by the average losing trade in $.

For example, suppose I flip a fair coin. Thus, the reliability of the system (whether it comes up heads or tails) is 0.5. The rules of the game are you make twice the amount you risk when you win and you lose the amount you risk when you lose. Thus, B = 2. The question is "What percentage of my remaining equity should I risk on each run to produce the maximum rate of return?" Maximum % = 0.5 – [(1 – 0.5)/2] = 0.5 – (0.5/2) = 0.5 – 0.25 = 0.25 Thus, a maximum risk of 25% would yield the largest returns in this game. However, a Kelly criterion only works when you have two possibilities, not a bag full. Expectancy Divided by Worst-Case Loss

Expectancy/ Worst possible risk against you

Note that the Kelly or the Expectancy criteria gives you a maximum ceiling for your bet. In reality, your maximum risk size should be nowhere near either level. It should be way below them!!!! Both could overshoot the optimal risk size as determined by the spreadsheet you were shown. Neither of these calculations is safe—both result in a high probability of ruin!

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Optimal f Since the Kelly Criterion is really based on a fixed bet size (as in gambling), Ralph Vince has come up with the idea of Optimal f, which is applied to a string of wins and losses in which the bet size varies.

Ralph Vince: "For any given independent trial situation, where you have an edge (i.e., a positive mathematical expectation), there exists an optimal fixed fraction (f) between 0 and 1 as a divisor of your biggest loss to bet on each and every event to maximize your winnings. Most people think that the optimal fixed fraction is the percentage of your total stake to bet. This is absolutely false. Optimal f is the divisor of our biggest loss, the result of which we divide our total stake by to know how many bets to make or contracts to have on." Portfolio Management Formulas, p. 80.

Tharp: Optimal f gives you a larger bet size, so the Kelly Criterion is actually a more conservative number. Even though it is more conservative, it still results in a high probability of ruin and huge drawdowns. Optimal f is based on the largest sized loss you have had in your historical testing of your system. Thus, if you use optimal f and you have not yet had your largest loss (a likely possibility since the variable of price change may be infinite), you might be in for a tremendous loss trading at optimal f. You must calculate optimal f for each market and each market system pair. There is no easy formula for calculating optimal f. You must use a computer and test all possible values between 0.01 and 1.00 in 0.01 increments or use some form of iteration. This amounts to finding the f that produces the largest Terminal Wealth Relative (TWR) in the following formula:

TWR = Σi from 1N trades of [1 + (– trade i/biggest loss)] Ralph Vince’s definition of optimal f means he uses the highest average ending equity to define this.

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Strategies to Avoid Joe Ross System

Trade 5 contract units, but get out of 3 contracts when you are slightly ahead of all costs. Raise stops on remaining contracts to break-even.

Larry Williams’ Martingale Strategies (Definitive Guide to Trading, Part 2)

System Expectations Are Low: When the actual results of the system are out-of-whack with the expectations of the system, then you can start increasing your commitment. For example, if you have a 65% system, when you’ve had 13 losers out of the last 20 trades (i.e., 65% losses), then increase your commitment by stepping up one unit. If you normally trade 1 unit, you now trade 2 units and continue to trade 2 units until you have winners in 13 out of 20 trades. Minimal Martingale: If you have a system that is only right about 30% of the time, you might do the following. After every losing trade, increase your commitment on the next trade by 1 unit. If you normally trade 1 unit, then you’d move up to 2 units after a loss. If you have another loss, you add another unit so that you’re trading 3 contracts. If that trade wins, then you’d move back to 2 units. If the next trade wins, you’d go back to 1 unit and stay at that level until you had a loss.

Fixed Ratio Trading (as defined by Ryan Jones) While fixed ratio trading may have some merits, the way Ryan Jones presents it has some huge holes

• Risk of ruin is meaningless. • 10% risk is treated as everyday. • You can trade 1 S&P contract and 1 corn and treat them as equal. • Jones’ methods are for the little guy while Van Tharp only works with big traders.

Fixed ratio trading can be an excellent method if you fix your risk (i.e., define R), define your worst case loss in terms of a cumulative R, relate your increment factor (delta) to your worst case loss, and tie your INC factor (e.g., 1 unit, 5 units) to what you are comfortable losing. (See The Definitive Guide to Position Sizing Strategies

book).

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute

Your Notes: ___________________________________________________________________________ ___________________________________________________________________________ ___________________________________________________________________________ ___________________________________________________________________________ ___________________________________________________________________________ ___________________________________________________________________________ ___________________________________________________________________________ ___________________________________________________________________________ ___________________________________________________________________________ ___________________________________________________________________________

Copyri

ght (c

) 202

1 Van

Tharp

Insti

tute