Two traders can run the exact same strategy — same pairs, same win rate, same reward-to-risk — and one doubles their account while the other is forced out of the market within a year. The difference is almost never the strategy. It is how much of the account was at risk on each trade, and how that compounds when losses arrive in a row.
A win rate tells you how often something happens. It says nothing about how much you lose when it doesn't, or how many losses in a row you can absorb before the account is unusable. Those two questions are what "risk of ruin" measures.
The arithmetic is unforgiving because losses and gains are not symmetric. Lose 50% of an account and you need a 100% gain to get back to where you started. Lose 80% and you need a 400% gain. This is why a strategy with a perfectly respectable win rate can still end in a wiped account: the losing runs do damage that the winning runs cannot repair fast enough.
Start with $10,000 and lose 10% per trade for 10 consecutive losses:
$10,000 → $9,000 → $8,100 → $7,290 → $6,561 → $5,905 → $5,314 → $4,783 → $4,305 → $3,874
You are down 61%, and you now need a +158% return on the remaining $3,874 to get back to break-even. The same 10 losses at 2% risk per trade would have left you at $8,171 — down 18%, needing +22% to recover.
Three inputs decide how likely you are to hit a ruin threshold, and only one of them is about your strategy:
Expectancy — the average result per trade in R — combines the second and third:
Expectancy = (Win rate × Reward:Risk) − (1 − Win rate)
At a 40% win rate and 2R average winner: (0.40 × 2) − 0.60 = +0.20R per trade.
At a 40% win rate and 1.5R: (0.40 × 1.5) − 0.60 = 0.00R — a break-even system before costs, and a losing one after spread and commission.
Positive expectancy is necessary but not sufficient. A +0.2R system with 5% risk per trade has a realistic chance of ruin; the same system at 0.5% risk is extremely unlikely to ever hit a 50% drawdown threshold. Expectancy determines whether the game is worth playing. Position size determines whether you survive long enough to be paid.
At a 45% win rate, the chance of any single trade being a loss is 55%. The chance of three losses in a row at some point across 200 trades is not 0.55³ applied once — it is close to a certainty. Streak length grows with sample size, and traders almost universally underestimate it.
| Win rate | Loss rate | Chance a 3-trade stretch is all losses | Chance a 6-trade stretch is all losses | Expected longest losing streak in 500 trades |
|---|---|---|---|---|
| 40% | 60% | 21.6% | 4.7% | ~11 |
| 45% | 55% | 16.6% | 2.8% | ~9 |
| 50% | 50% | 12.5% | 1.6% | ~8 |
| 60% | 40% | 6.4% | 0.4% | ~6 |
The first three columns describe any given stretch of trades; the last column is the one professionals plan around — over a few hundred trades, the longest losing run you should expect to experience. Read it again: even at a 60% win rate, over 500 trades you should expect to live through a run of roughly six consecutive losses. At 1% risk per trade that is a 5.9% drawdown — uncomfortable but survivable. At 5% risk per trade it is a 26% drawdown, and at 10% it is roughly a 47% drawdown that will very likely breach a prop-firm limit or an emotional breaking point.
The Kelly criterion gives the position size that maximises long-run growth for a known edge:
f* = Win rate − (1 − Win rate) ⁄ Reward:Risk
At 45% win rate and 2R: f* = 0.45 − (0.55 ⁄ 2) = 0.175 — 17.5% of capital per trade.
That number is almost always far too aggressive in practice. Full Kelly assumes you know your edge exactly, that trade outcomes are independent, and that you can tolerate the drawdowns it produces — which routinely exceed 50%. Real edges are estimated from limited samples, real losses cluster through correlation and volatility regimes, and real traders stop following their plan long before the maths says they should. Most professional risk practice sits between a quarter and a half of Kelly, and many desks cap risk per trade outright regardless of what Kelly says.
Forextity's Risk of Ruin simulator runs thousands of simulated accounts through your inputs and reports the distribution, not a single outcome:
The simulation is deterministic: the same inputs always produce the same numbers, so a result you write down can be reproduced and audited. It is also deliberately simple — each trade is an independent draw. Real trading is worse in one specific way: losing trades arrive clustered, correlated to volatility, and often right after the decision-making has already degraded. Treat the simulation as a floor on how bad things can get, not a ceiling.
See your own numbers. Run your win rate and reward:risk through thousands of simulated accounts and check the ruin probability before you size the next trade.
Open the Risk of Ruin Simulator →