Risk of ruin is the probability that an account draws down to a predefined failure point — a hard stop-out, a prop-firm drawdown limit, or simply the level of pain at which you stop trading the system — before its positive edge has time to compound. It is the single number that connects your edge (win rate and payoff ratio) to your bet sizing (risk per trade) and tells you whether a profitable strategy can actually be traded through a normal losing streak without blowing up first.
This calculator does not use the closed-form gambler’s-ruin shortcut, which assumes every win and loss is the same fixed size. Instead it runs a Monte Carlo simulation — 1,500 independent account paths of 800 trades each — drawing wins and losses at random according to your inputs and counting the fraction of paths that touch the ruin threshold at any point along the way. Because it samples the full sequence of outcomes rather than a single average, it captures the path dependency that destroys real accounts: it is the order of the losses, not just their count, that ruins you.
Why risk of ruin matters
A positive expectancy is necessary but not sufficient. Two traders can run the identical edge and one survives while the other is wiped out, purely because of how much they risked per trade and the sequence in which variance arrived. Risk of ruin is the bridge between "this system makes money on average" and "this account is still open in six months." It reframes sizing as a survival question rather than a return-maximizing one.
The relationship between risk per trade and ruin is brutally non-linear. Doubling your fixed risk from 1% to 2% does not double your odds of ruin — it can multiply them by an order of magnitude, because deeper individual losses compound into drawdowns that the same edge can no longer recover from. Seeing that curve is what turns "risk small" from a platitude into a specific, defensible number for your particular edge.
For prop traders the ruin threshold is not abstract — it is the firm’s maximum drawdown, and breaching it ends the account regardless of long-run expectancy. Running this simulation before you accept a sizing scheme tells you whether the challenge or funded account is mathematically survivable at your edge, or whether you are being asked to thread a needle that variance will not allow.
The formula
RoR ≈ (number of simulated paths that hit the ruin threshold) / N, where N = 1,500 paths × 800 trades each
RoRRisk of ruin — the estimated probability (0–100%) that an account reaches the ruin threshold within the simulated horizon.
NNumber of independent simulated account paths. Here N = 1,500; each path is an 800-trade sequence.
WWin rate — the probability any single trade is a winner, entered as a percentage.
RPayoff ratio (reward-to-risk) — the size of the average win expressed in multiples of the average loss (the R-multiple of a winner).
fRisk per trade (fraction risked) — the share of current equity wagered on each trade, e.g. 0.01 for 1%.
DRuin threshold — the drawdown from peak (or absolute equity floor) that counts as ruin, e.g. a 50% drawdown or a prop-firm max loss.
Each simulated trade compounds on current equity: a win adds f × R × equity, a loss subtracts f × equity. A path is counted as ruined the first moment its equity touches D; it is not double-counted if it recovers. Because the result is a sample estimate, re-running produces a small statistical wobble (roughly ±1–2 percentage points), which shrinks as paths increase.
Worked example: a 45% win rate at 2R
- · Win rate W = 45% (you win just under half your trades).
- · Payoff ratio R = 2.0 (winners average twice the size of losers).
- · Risk per trade f = 2% of current equity, fixed-fractional.
- · Ruin threshold D = a 50% drawdown from the starting balance.
- · Horizon = 800 trades, simulated across 1,500 independent paths.
Expectancy per trade (in R)(0.45 × 2) − (0.55 × 1) = 0.90 − 0.55 = +0.35R
Edge per trade as % of equity0.35R × 2% = +0.70% expected growth per trade — a genuinely positive system
Single-path mechanicsEach win multiplies equity by 1.04 (1 + 0.02×2); each loss multiplies it by 0.98 (1 − 0.02)
Simulate 1,500 paths × 800 tradesDraw each trade as a win with probability 0.45, flag any path whose equity ever falls to 0.50 of start
Count ruined pathsRoughly 90 of 1,500 paths breach the 50% drawdown at some point → RoR ≈ 6%
Even a clearly profitable +0.35R edge carries a meaningful ruin risk at 2% per trade — cut risk to 1% and that same edge typically drops ruin to a fraction of a percent.
How to use it
- 01Enter your win rate — the percentage of trades that close as winners over a meaningful sample (ideally 100+ trades, not last week).
- 02Enter your payoff ratio (average win ÷ average loss), expressed as an R-multiple such as 1.5 or 2.0.
- 03Set your risk per trade as a percentage of equity — the amount you actually lose when a stop is hit, not your intended target.
- 04Define the ruin threshold: the drawdown or equity floor that ends the game for you (a prop max loss, or a personal line like 40–50%).
- 05Run the simulation and read the ruin probability, then re-run it once or twice to see the statistical spread.
- 06Adjust risk per trade up and down to map the trade-off, and pick the sizing where ruin sits comfortably inside your tolerance.
Common mistakes
Treating a low number as zero
A 3% risk of ruin is not "safe" — it means roughly 1 in 33 traders running this exact edge gets wiped out. Across a long career and multiple accounts, small per-run probabilities accumulate.
Using optimistic, in-sample stats
Plugging in the win rate and payoff from your best month, or from a curve-fit backtest, understates ruin badly. Feed it conservative, out-of-sample numbers — ruin is hypersensitive to a few points of win rate.
Ignoring path dependency
Traders reason "I have a positive expectancy, so I cannot be ruined." Expectancy is an average over infinite trades; ruin happens in the finite, unlucky sequences where losses cluster before the edge can compound.
Risking a fixed dollar amount, not a fraction
Fixed-dollar risk raises your effective percentage as the account draws down, accelerating ruin exactly when you are most vulnerable. Fixed-fractional sizing, which this model assumes, automatically shrinks bets in a drawdown.
FAQ
Why use a Monte Carlo simulation instead of the classic risk-of-ruin formula?
The textbook formula assumes every win and loss is identical in size and ignores compounding and a finite horizon. Real edges have a payoff ratio above or below 1 and real accounts compound. Simulating 1,500 full sequences captures the variance and path dependency that the closed form averages away.
What counts as "ruin"?
Whatever level ends your ability to keep trading the system: a prop firm’s maximum drawdown, a broker margin stop-out, or a personal drawdown threshold past which you would abandon the strategy. The number is only as meaningful as the threshold you choose, so set it honestly.
My result changes slightly each time I run it. Is that a bug?
No — it is sampling noise. A Monte Carlo estimate from 1,500 paths carries a confidence interval of roughly ±1–2 percentage points. Treat the output as a range, not a decimal-precise constant, and focus on the order of magnitude.
How do I lower my risk of ruin?
The fastest lever is risk per trade — because ruin scales non-linearly, halving your fractional risk usually cuts ruin by far more than half. Beyond that, improving win rate or payoff ratio raises expectancy and pushes ruin down, but sizing dominates in the short to medium term.
Does a positive expectancy guarantee I will not be ruined?
No. Positive expectancy guarantees profit only in the limit of infinite trades with unlimited capital. Over a finite 800-trade horizon with a finite balance, an unlucky early cluster of losses can hit the ruin threshold before the edge ever compounds in your favor.
What is an acceptable risk of ruin?
There is no universal number, but most disciplined traders target well under 1% for a meaningful drawdown threshold. If your inputs produce double-digit ruin, the edge is either too thin or, far more often, the position size is too large.