Position Sizing Math and Risk of Ruin for Futures Traders

Position size decides whether a normal losing streak is an annoyance or the end of the account. Most traders pick size from what they hope to make. A better starting point is the loss that stops you, and how likely it is that your strategy hits it. That probability is called risk of ruin.

What risk of ruin means

Risk of ruin is the chance that your equity falls from where it is now to a level that ends the game. For a personal account that is a threshold you choose, such as the point where you stop trading. For a prop evaluation it is the drawdown floor. The math answers a narrow question: given outcomes like these, at this bet size, how often does the path touch that floor?

The simple formula and its assumptions

The classic version applies to a repeated bet where every trade either wins or loses exactly one unit of risk (call it 1R), trades are independent, and the win probability p stays constant. With p above 0.5, the chance of ever losing N units is:

Risk of ruin = ((1 - p) / p) to the power N

This is the same as ((1 - edge) / (1 + edge)) to the power N, where edge = 2p - 1. N is the number of 1R losses between you and the floor, meaning your buffer divided by your risk per trade.

Everything about real trading breaks those assumptions a little. Winners and losers are not the same size. Risk per trade changes. Stops slip. Prop drawdowns trail your equity. For fixed outcomes, the formula is a useful estimate. For variable dollar risk, trailing drawdown or time limits, you need a simulation that samples your actual trade results. Never treat the simple number as a promise.

Hypothetical example

Assume a strategy that wins 55 percent of the time with equal winners and losers of 1R. This is an assumption for illustration, not a result from any strategy.

The numbers come from (0.45 / 0.55) to the N. Same strategy, same win rate, and the ruin probability drops from roughly 1 in 7 to 1 in 400 just by risking a third as much per trade. With a $3,000 buffer, N = 10 means $300 risked per trade, and N = 30 means $100. Notice also what happens at 50 percent with equal sizes: edge is zero, and with no edge ruin is certain over enough trades. Costs make a zero-edge strategy negative.

Losing streaks are a separate calculation

Another way to look at it: how many consecutive full losses can you absorb? Hypothetical $3,000 buffer:

The third column is the chance that a specific run of trades is all losses. It is not the chance you meet such a streak somewhere in a long sequence. Over 100 trades at a 50 percent loss rate, a run of six losses somewhere is more likely than not. This is why a buffer of six trades is thin even when the single-sequence figure looks small. It is also why simulating your own sequence gives a better answer than multiplying probabilities.

Converting percent risk into contracts

Contract values: ES is $50 per point ($12.50 per tick), MES is $5 per point ($1.25 per tick), NQ is $20 per point ($5 per tick), and MNQ is $2 per point ($0.50 per tick). Margin is collateral, not a cap on loss. A stop order can fill past its price, so size from your stop distance with some room.

Worked example, $10,000 account at 1 percent risk, which is $100. A 4-point MES stop is 16 ticks at $1.25, which is $20 per contract. Five contracts gives $100 of planned risk. One ES contract with the same 4-point stop is $200, which is 2 percent, so MES gives you the finer step.

On a prop account the percent that matters is the percent of the drawdown buffer, not the nominal account. With a hypothetical $3,000 buffer, $30 per trade is 1 percent, which is very conservative. Whatever you choose, trailing rules differ between firms, so read the exact rule before you decide how much of the buffer a single trade may use.

Building your own sizing model

1. Log your trades with real or realistically assumed costs. A hundred trades is the least you should have before trusting any estimate, and it is still a small sample.
2. Compute win rate and average winner and loser. With 100 trades and a 50 percent observed win rate, the standard error is about 5 percentage points, so the true rate could plausibly sit anywhere in a wide band.
3. Decide your floor, either the prop drawdown or your personal stop level.
4. Divide the floor by your planned risk per trade to get N.
5. Run a simulation that resamples your trade results (or at least test pessimistic win rates) to see how often the floor is touched, including trailing rules if they apply.
6. Pick the size where the result is acceptable at your pessimistic inputs, not your best ones.

The edge calculator is a quick way to check expectancy before you size anything. If expectancy does not hold up with costs included, no position size fixes it.