Risk of ruin and account size: how much to risk to survive

8 min read

Having a strategy with an edge does not protect you from blowing up your account: that depends on how much you risk on each trade. Risk of ruin is the probability of losing so much of your capital that trading is no longer viable. This guide explains how it is calculated, why risking 1% or 5% changes the outcome so radically, and which practical limits are worth setting.

What risk of ruin is

Risk of ruin is the probability that your account falls to a level you consider unacceptable (for example, 50% of starting capital, or the whole balance) before your strategy’s edge has time to show itself. It matters because losses are asymmetric: the deeper the drop, the harder it is to recover, and a strategy that is profitable in the long run can be wiped out along the way.

Three things determine it: the win rate, the relative size of wins and losses (the R:R) and, above all, the fraction of the account you risk per trade. Of the three, only the last is entirely under your control.

Losing streaks: exact figures

If each trade is independent and you lose with probability q = 1 − WR, the probability of n losses in a row is q^n. With a 50% win rate: three losses in a row, 12.5%; five, 3.125% (one in 32); ten, 0.0977% (one in 1,024). With a 40% win rate (q = 0.6): five, 7.78%; eight, 1.68%; ten, 0.605%; fifteen, 0.047%.

What matters is that these probabilities apply to one specific block of trades. When you make hundreds of trades, there are many possible blocks. Simulating 200 trades with a 40% win rate, the longest losing streak averages about 9, and the probability of seeing at least one streak of 10 or more is around 38%. With a 50% win rate, the average longest streak in 200 trades is about 7. Anyone who says “I never lost ten in a row” has a small sample.

What a streak costs at your risk level

The effect of a streak depends on how much you risk. With compounding losses, n losses in a row at risk r leave the account at (1 − r)^n. Ten losses in a row at 1% per trade leave it at 0.99^10 ≈ 0.904, a drop of 9.56%. At 2%, it is 18.29%. At 5%, 0.95^10 ≈ 0.599, a drop of 40.13%.

And recovering takes more than you lost: after −9.56% you need +10.57%; after −18.29%, +22.39%; after −40.13%, +67.02%. After ten losses in a row at 5% you are in a position where starting over is hard both psychologically and mathematically. You can check these figures in the drawdown and recovery calculator.

The classic gambler’s formula

When you always risk the same fixed amount (one unit of risk) and wins and losses are symmetric, the probability of total ruin starting from N units is (q ÷ p)^N, where p = win rate and q = 1 − p, provided p > q. With a 55% win rate (q ÷ p = 0.45 ÷ 0.55 ≈ 0.818): if you risk 10% of starting capital on each trade (N = 10), ruin is 13.4%; at 5% (N = 20), 1.8%; at 1% (N = 100), about 2 × 10^−9, which is practically zero.

With 2 R wins and a 40% win rate (expectancy +0.2 R), the equation solves differently, and the probability of ruin is 0.8229^N: 14.2% for N = 10, 2.0% for N = 20 and 3 × 10^−9 for N = 100. This model simplifies because it treats risk as a fixed amount rather than a percentage of an account that changes, but it shows the essence: cutting your risk in half does not halve the probability of ruin, it reduces it by far more.

Monte Carlo simulation explained in words

In practice risk compounds with the account, and there is no convenient formula for every case. That is why Monte Carlo simulation is used: the computer generates thousands of possible “lives” of your account. In each one it flips a weighted coin for every trade (with a probability equal to your win rate), adds or subtracts the win or loss in R multiplied by your risk percentage, and records whether the balance ever reached the ruin level you defined.

The share of “lives” that end in ruin is the estimated risk of ruin. The more repetitions, the more stable the estimate. Our risk of ruin simulator runs this calculation in your browser with your own strategy figures, without sending them to any server.

1% versus 5%: a simulated example

Take a strategy with a 45% win rate, a 1.5 R win and a 1 R loss (expectancy +0.125 R), 1,000 trades, and “ruin” defined as touching 50% of starting capital. In a simulation of tens of thousands of runs, the approximate results were: risking 1% per trade, under 0.1% ruin; 2%, around 0.5%; 3%, around 4%; 5%, around 19%; and 10%, around 60%. Your own simulation may vary by a few tenths with a different random seed, but the pattern does not change.

The strategy is identical in every case. The only thing that changes is position size. Moving from 1% to 5% turns a practically zero risk into almost a one-in-five chance, while the expected profit per trade only grows fivefold. A strategy with an edge can therefore blow up through over-leverage.

Expected drawdown

In that same simulation, the median maximum drawdown (the fall from peak) was about 19% at 1% risk, 35% at 2%, 49% at 3%, 70% at 5% and 95% at 10%. At the 95th percentile (a bad but plausible scenario) it was roughly 30% at 1%, 52% at 2% and 88% at 5%.

When you decide how much to risk, do not look only at the average drawdown: look at the plausible bad case and ask whether you would keep trading with discipline through that drop. If a 30% loss would make you abandon your rules, your real risk is lower than the numbers suggest.

Without an edge, ruin is a matter of time

With negative or zero expectancy, risk of ruin tends toward 1 if you trade long enough; lowering risk only delays it. In the same simulation, with a 45% win rate and a 1 R win (expectancy −0.10 R), 1,000 trades wipe out 90% of accounts at 1% risk, 99.6% at 2% and nearly all at 5%.

In other words, risk management does not turn a losing strategy into a profitable one. It slows the rate of loss and protects you when there is an edge. That is why expectancy (see the guide on expectancy and profit factor) and risk of ruin should be analyzed together.

Account size and practical limits

Account size does not change risk of ruin in percentage terms, but it does change practical limits. With 500 USD and 1% risk (5 USD), a 25 pip stop on EUR/USD, where a micro lot (0.01 lots) is worth 0.10 USD per pip, allows 5 ÷ (25 × 0.10) = 2 micro lots, that is 0.02 lots. With a 50 pip stop, only 0.01 lots. With 100 pips, the minimum 0.01 lots already risks 10 USD, or 2%. A small account forces you into tight stops or into risking more than planned.

Set limits: maximum risk per trade of 0.5% to 2%; a maximum daily loss (for example, 3%) and a weekly one; reduced size after a drawdown; correlated positions counted as one; and an allowance for the fact that on gaps and news your stop can fill worse than planned. Work out size with our lot size calculator rather than deciding by eye.

Common mistakes

Raising risk to win back losses (the fastest way to speed up ruin); using the win rate from a short backtest as if it were the true one; leaving costs out of the simulation; assuming trades are independent when you open several correlated positions; and defining ruin as “losing everything” when in reality you would stop trading long before that.

Summary and practical steps

The probability of n losses in a row is (1 − WR)^n, and with hundreds of trades long streaks are normal. Risk per trade dominates risk of ruin: going from 1% to 5% can turn a near-zero probability into a double-digit one. Decide what drawdown you could tolerate, run several risk percentages (conservative ones) with your own figures in the risk of ruin simulator, set daily and weekly limits, reduce size after a bad streak and log everything in your journal. The goal is to still be trading when the bad streak ends, because it will come.

Educational content, not financial advice. Risk disclosure

Related tool

Risk of ruin simulatorSimulate thousands of trade sequences (Monte Carlo) to estimate the probability of blowing up your account based on your system and risk per trade.

More guides