Max drawdown position sizing: how many contracts fit your limit

Contents

1. How do you size futures positions to a max drawdown limit?
2. Why is max drawdown the right sizing input, not margin?
3. How much max drawdown limit should you set?
4. How do you use drawdown to decide the number of contracts?
5. How much money do you need to run how many contracts?
6. Why add a Monte Carlo buffer, and when do micro contracts help?
7. Frequently asked questions
8. References

How do you size futures positions to a max drawdown limit?

Max drawdown position sizing means dividing the drawdown you can accept by the drawdown one contract produced in the backtest, after widening that figure with a buffer, and rounding down to whole contracts. Margin tells you what the exchange allows. Max drawdown tells you how many contracts you can hold through the worst stretch.

Max drawdown is the largest fall from an equity peak to the following low, measured before the equity makes a new high. It is the deepest hole the strategy dug on the way to its final result. If you already have an equity series from a backtest, the free drawdown calculator reads that hole off the curve for you.

The reason drawdown is the right input is simple arithmetic. If one contract produced a certain drawdown on a set of trades, two contracts on the same trades produce twice that drawdown in dollars, three contracts three times, and so on. Think of stacking boxes under a ceiling: each box is the one-contract drawdown, the ceiling is the drawdown you have decided you can take, and the number of boxes that fit under the ceiling is your contract count. The ceiling, not the exchange margin table, should decide your size. That is the whole argument of this article.

Why is max drawdown the right sizing input, not margin?

Margin and max drawdown answer different questions, and only max drawdown answers the question of survival. Margin is the deposit the exchange requires you to hold for each open contract. It is set by the exchange from recent price movement and says nothing about how long a losing streak your strategy can produce or how deep it goes. A trader who fills the account with as many contracts as the margin allows has answered "what is permitted" and skipped "what can I live through".

The one-contract max drawdown from your own backtest is the number that scales directly with contract count. It comes from the strategy's own trades, in the order they happened, with your own costs included. Where the losses behind a drawdown come from, a single large loss or a run of ordinary losses in a row, decides which exits you test next; that question is covered in stop loss in a futures strategy: fixed ticks, ATR, time or trailing.

Question	Size by margin	Size by drawdown
What the number measures	The exchange's deposit per contract	The worst losing stretch of your strategy
Who sets it	The exchange	Your backtest and your written limit
What happens when it is wrong	A margin call and forced reduction	The account falls below a limit you chose in advance
Typical contracts on the same account	More	Fewer

The run below is the one-contract backtest this article sizes from. It tested a Stochastics rule as its only entry and exit, no stop or target, no filters, on NQ daily bars from 2019-01-02 to 2025-06-30, long and short, with $20 commission and $10 slippage per trade. It is a stand-in rule built from standard indicators, not a sizing method. The search in AlphaPrime (Valynox) tried about a thousand candidates and kept this one for its return against its drawdown; picking the kept one after the fact flatters it, and the rule did not earn this money for anyone.

The number this article uses from the box is the max drawdown of $32,935 per contract. On the equity curve it is the deepest dip from a peak to the low that follows, and that dip is the sizing input.

How much max drawdown limit should you set?

Set the limit as the fraction of the account you could lose and still keep trading the system without changing it, and write it down before you size. A drawdown limit is the loss from an equity peak at which you have decided, in advance, to act. The test for the right fraction is behaviour, not mathematics: at what loss would you start skipping signals, halving size by feel, or switching to a different strategy? The limit must sit above that point, because a system that is abandoned in its drawdown never reaches its recovery.

Three zones make the choice concrete. Above the written limit you keep trading as planned. Below it you review the system against its backtest, because the live result has gone outside what you expected. Near the bottom of the account you stop, because the capital left can no longer support one contract through another drawdown. Many systematic traders write the limit somewhere between a fifth and a third of the account; a smaller fraction means fewer contracts and a longer trading life, a larger fraction means more contracts and a shorter one.

Expect the live drawdown to run deeper than the backtest figure, for two separate reasons. First, the backtest result was the best of many candidates, so its drawdown is flattered by selection. Second, the longer a strategy runs, the more chances it has to produce a new deepest dip; Magdon-Ismail and Atiya (2004) show that the expected maximum drawdown of a random process keeps growing as the measured period lengthens. A live account that runs longer than the backtest should be expected to show a deeper hole, which is why the next section widens the backtest figure before dividing.

How do you use drawdown to decide the number of contracts?

Four ordered steps turn a one-contract backtest into a whole-contract count, and the order matters because the buffer must be applied before the division.

1. Take the one-contract max drawdown from the backtest. Use the run with your real commission and slippage, long and short as you will trade it.
2. Widen it with a buffer. Multiply by a factor above one. The last section explains where the factor comes from; a Monte Carlo reshuffle of the trade order gives a measured basis for it.
3. Divide your drawdown limit in dollars by the widened figure. The limit in dollars is the account size multiplied by the limit fraction from the previous section.
4. Round down to whole contracts. Rounding up puts the buffered drawdown above the limit by construction. If the result is below one, the strategy does not fit this account at this limit.

Every number in a backtest report that is measured in dollars moves with the contract count: net profit, average trade, max drawdown and the size of the largest single loss. Ratios such as profit factor and win rate do not. What changes and what stays fixed when you trade more than one contract is the subject of futures lot size and how contract count changes a backtest.

How much money do you need to run how many contracts?

The account size follows from the contract count, not the other way round. Most traders start with the money they have and ask how many contracts it supports; the better question is how many contracts the strategy and the limit justify, and what account that needs. The one formula in this article rearranges the division of the previous section:

account = contracts × buffered one-contract drawdown ÷ limit fraction

Here contracts is the number you want to trade, buffered one-contract drawdown is the backtest drawdown after the buffer from step two, and limit fraction is the share of the account you wrote down as your drawdown limit, as a decimal (30 percent is 0.30).

The second box shows what sizing alone does to the same trades. AlphaPrime ran the 62 trades from the box above under nine money-management methods, with capital of $100,000, at most 10 contracts, and each method's report defaults. The entries and exits are identical in every row; only the number of contracts per trade changes, so every difference in the table comes from sizing.

Read the drawdown column against the contracts column. Kelly averaged 1 contract and reached a max drawdown of $32,935, the same figures as the one-contract box, because in this run the method held one contract throughout. Fixed fractional risk averaged 4.26 contracts and reached a max drawdown of $143,320, more than four times as deep, and its net profit of $78,800 is lower than the $150,620 of Fixed contracts at 3, because the extra contracts were largest at the wrong moments. A drawdown larger than the starting capital can appear in the table because each drawdown is measured from an equity peak that was higher than the start. The Kelly row is the fully invested end of the scale in theory, and why most traders run a fraction of it is explained in Kelly criterion position sizing on futures: full, half or fixed risk.

Why add a Monte Carlo buffer, and when do micro contracts help?

One backtest drawdown is one ordering of the trades, so the buffer should come from the spread of drawdowns that other orderings would have produced. A Monte Carlo buffer is the extra width you add to the backtest drawdown after reshuffling the trade order many times and reading the max drawdown of each reshuffle. The method takes its name from the random sampling approach described by Metropolis and Ulam (1949). The everyday picture is a deck of cards: the same cards dealt in a different order can put four losing hands together that the first deal spread out, and a losing streak that lands together digs a deeper hole than the same losses spread apart.

The reshuffles give a distribution of max drawdowns rather than one number. The backtest figure sits somewhere inside it, often below the middle, because the historical order was one of the kinder ones for a candidate that was kept for its return against its drawdown. You size to the upper part of that spread: pick a level in advance, for example the drawdown that nine reshuffles in ten stay under, and use it as the buffered figure in step two. The ratio between that level and the backtest drawdown is your buffer factor. AlphaPrime shows this distribution in the Performance Report, on the Monte Carlo screen next to the equity curve and trade list, so the upper figure is read from the same report as the one-contract drawdown. If your trades come from another source, the free Monte Carlo simulator builds the same distribution from a pasted list of trade results.

Volatility changes the buffer too. When the market's daily range widens, the losses per trade widen with it even if the rule is unchanged, so a drawdown measured in a quiet period understates what the same rule does in a loud one. Reading the stop distance and the position from recent range, as in ATR trailing stop and position size on futures, keeps the per-trade loss closer to a fixed dollar amount and makes the Monte Carlo spread narrower.

Micro contracts help when the division lands between whole contracts. On the CME, the E-mini Nasdaq-100 future (NQ) is $20 per index point and the Micro E-mini Nasdaq-100 future (MNQ) is $2 per index point, so ten MNQ equal one NQ in point value. Exchanges can change contract specifications, so check the exchange's contract page; the free futures contract calculator shows the current tick and point values side by side. With micros, a computed size of 1.7 NQ becomes 1 NQ plus 7 MNQ instead of a forced choice between 1 and 2 NQ. Check your broker's per-contract commission, because the cost per point of exposure can differ between the two sizes.

The next thing to do is to open your own one-contract report, write down its max drawdown and the upper Monte Carlo figure, choose your limit fraction before looking at either, and do the division. Then compare the contract count you get with the count your margin would have allowed, and note the difference.

Frequently asked questions

What is a reasonable max drawdown limit for a futures account?

A limit is reasonable if you would keep trading the system unchanged while the account is above it. Many systematic traders write it somewhere between a fifth and a third of the account; a smaller fraction gives fewer contracts and a longer trading life. Write the fraction down before sizing, and size to a buffered drawdown, not the raw backtest figure.

How much money do I need to trade two contracts?

Multiply the buffered one-contract max drawdown by two and divide by your limit fraction. The buffered drawdown is the backtest figure widened by a Monte Carlo reshuffle of the trade order. The account size that comes out is the one that keeps the two-contract drawdown under your limit; margin is usually a smaller number and is not the constraint.

Can I size positions by margin instead of drawdown?

Margin tells you the largest position the exchange permits, not the largest position you can hold through a losing streak. Two traders with the same margin and different strategies face very different drawdowns per contract. Size by the drawdown of your own backtest, then confirm that the resulting position also fits within margin.

Do I need Monte Carlo to size positions?

You can size without it, but then the only drawdown you have is the one ordering of trades that history happened to produce. Reshuffling the trade order shows how deep the same trades can go in a less friendly order, and the upper part of that spread is a more honest sizing input. The reshuffle needs only the list of trade results, which any backtest report provides.

References

• Magdon-Ismail, M., & Atiya, A. F. (2004). Maximum Drawdown. Risk, 17(10), 99–102.
• Metropolis, N., & Ulam, S. (1949). The Monte Carlo Method. Journal of the American Statistical Association, 44(247), 335–341.
• CME Group. E-mini Nasdaq-100 Futures Contract Specs. https://www.cmegroup.com/markets/equities/nasdaq/e-mini-nasdaq-100.contractSpecs.html
• CME Group. Micro E-mini Nasdaq-100 Index Futures Contract Specs. https://www.cmegroup.com/markets/equities/nasdaq/micro-e-mini-nasdaq-100.contractSpecs.html