Trading Risk of Ruin Calculator
Strategy and risk assumptions
Finite-horizon ruin risk
“Ruin” here means touching the reader-chosen floor relative to starting equity. It is not literal bankruptcy and it is not a peak-to-trough maximum-drawdown calculation.
How ruin risk builds across the trade horizon
These checkpoints are cumulative first-passage probabilities. A path that crossed the floor earlier is removed from the surviving state before later trades are evaluated, so it is counted exactly once.
| Through trade | Cumulative ruin probability | Survival probability |
|---|---|---|
| 1 | 0% | 100% |
| 25 | 2.918% | 97.082% |
| 50 | 5.039% | 94.961% |
| 75 | 5.921% | 94.079% |
| 100 | 6.384% | 93.616% |
What the entered edge implies
At a 1:1 average win/loss ratio, the shared payoff model has a 50% break-even win rate. The entered assumptions imply +0.200 of modeled arithmetic expectancy per $1 risked and +0.876% expected log growth per trade at the entered fixed-fraction risk.
Positive expectancy does not eliminate path risk. Even a favorable repeated process can cross a severe equity-loss threshold before its long-run edge has time to dominate. That distinction is why this tool measures the path event directly instead of inferring safety from win rate or Kelly alone.
How the exact calculation works
A modeled win multiplies then-current equity by 1 + fR, while a modeled loss multiplies it by 1 − f, where f is the entered equity-risk fraction and R is average win divided by average loss. After a given number of trades, equity is determined by the counts of wins and losses.
The dynamic program advances the probability of every still-surviving win/loss-count state one trade at a time. Any transition that reaches or falls below the starting-equity floor is moved permanently into the ruin probability. This produces an exact finite-state result under the stated binary assumptions—no random sampling or Monte Carlo seed is involved.
Use Kelly and losing-streak analysis for different questions
The Kelly Criterion Calculator asks what fixed risk fraction maximizes expected log growth under the same simplified binary payoff assumptions. This calculator instead asks how likely a chosen risk fraction is to cross a chosen starting-equity floor over a finite horizon.
The Trading Losing Streak Calculator measures the chance of a particular run of losses anywhere in a sample. A ruin event can occur through many different win/loss sequences, so a losing-streak probability is not a substitute for this first-passage calculation. The Backtest Win Rate Confidence tool can help stress the uncertainty in the win-rate estimate itself.
Model boundaries
The calculator assumes independent repeated opportunities, one constant win probability, one constant average win/loss ratio, and the same fixed fraction of then-current equity at risk on every trade. It does not model correlated trades, regime changes, variable payoffs, partial wins or losses, gaps beyond the planned loss, simultaneous positions, leverage or margin constraints, commissions, spreads, slippage, taxes, or parameter-estimation uncertainty.
The threshold is measured from starting equity, not from a rolling high-water mark. Results are conditional model outputs, not forecasts or recommended risk limits. Small errors in win rate, payoff ratio, or realized loss size can materially change the true path risk.
Method context
- Stanford: Risk-Constrained Kelly Gambling, discussing wealth-growth optimization alongside the probability of falling to an undesirable wealth level.
- Financial Analysts Journal: Liquidity Requirements for Financial Futures Investments, an example of finite-horizon probability-of-ruin analysis in financial risk management.