Risk of ruin is the probability that a trading process loses enough capital to become unusable before it reaches its objective. “Ruin” does not have to mean the account reaches zero. A strategy can be operationally ruined when equity falls below a minimum viable level, a margin threshold, a drawdown limit or the amount of capital needed to trade the strategy as designed.
The idea matters because a strategy can have a positive long-run expectation and still be traded so aggressively that normal variance creates an unacceptable chance of failure before that expectation has time to emerge.
Define what ruin means before trying to measure it
A useful risk-of-ruin analysis begins with a threshold. For one trader, ruin may mean a 30% drawdown because the mandate requires trading to stop there. For another, it may mean equity falling below the minimum size needed to execute the strategy efficiently.
Without a stated threshold, “risk of ruin” becomes vague. The threshold should reflect the actual capital objective and operating constraints, not just the mathematical possibility of a zero balance.
Risk per trade is one of the strongest drivers
The larger the fraction of capital exposed to each decision, the faster a losing sequence can push the account toward the ruin threshold. This is why risk per trade cannot be chosen in isolation from loss frequency and drawdown tolerance.
Reducing position size does not make the strategy more accurate. It gives the strategy more opportunities to survive ordinary variance and adverse sequences before the capital base becomes unusable.
Expectancy matters, but so does the distribution of outcomes
A process with negative expectancy tends to erode capital over repeated decisions and increases the probability of eventually crossing a serious loss or failure threshold. But a positive average does not eliminate ruin risk because results arrive in sequences, not as a smooth average.
Two strategies can have similar average returns while producing very different loss streaks, tail losses and drawdowns. Risk-of-ruin analysis therefore needs more than a win rate or a single reward-to-risk ratio.
There is no universal risk-of-ruin formula
Simple formulas can be useful under restrictive assumptions, such as independent trades with fixed win and loss amounts. Real trading often violates those assumptions. Position size may change, outcomes can be skewed, losses can gap beyond stops, and several positions may depend on the same market driver.
For that reason, MFXG does not present one closed-form formula as a universal answer. The method should match the strategy's actual payoff structure and the evidence available.
Correlation can make independent-looking trades fail together
If several positions share the same currency, index, volatility regime or macro driver, treating them as independent can understate the chance of a large combined loss.
The correlation risk framework therefore belongs inside ruin analysis. A portfolio with ten positions is not necessarily ten separate opportunities if one market shock can damage most of them at the same time.
Leverage can shorten the path to the threshold
Leverage increases market exposure relative to the capital supporting it. When the market moves adversely, the equity impact can therefore be much larger than the underlying percentage move suggests.
Margin requirements can also create forced actions before the trader reaches a self-defined loss limit. The Leverage & Margin guide explains why broker or clearing requirements are operational constraints, not substitutes for a risk budget.
Drawdown and ruin are related but not the same concept
Drawdown measures an observed decline from a previous equity peak. Risk of ruin asks a forward-looking question: given the strategy, position sizing and uncertainty, what is the chance that losses eventually cross a failure threshold?
A trader can be in drawdown without being close to ruin. Conversely, a highly leveraged account can be only a few adverse moves away from an operational threshold even if its current drawdown looks modest.
Simulation can expose sequence risk
When a strategy has a usable history of trade outcomes, resampling or simulation can help show how different orderings of wins and losses affect drawdown and threshold breaches. The purpose is not to predict the exact future path. It is to observe how sensitive the capital is to plausible sequencing.
Any simulation is only as good as its assumptions. If it assumes future losses cannot exceed historical losses, ignores changing correlation or treats every trade as independent, the resulting ruin probability can be falsely comforting.
Stress scenarios should challenge the model
A useful ruin review also asks what happens outside the average case: a longer losing streak, worse slippage, a volatility jump, simultaneous losses across correlated positions or a period when the strategy's edge weakens.
The scenario and stress testing page owns that broader process. Here the point is simple: a ruin estimate should not rely only on the most comfortable assumptions.
Common risk-of-ruin mistakes
- defining ruin only as an account balance of zero;
- using a formula whose assumptions do not match the strategy;
- treating historical win rate as a guaranteed future probability;
- ignoring correlated positions and leverage;
- assuming the worst historical loss is the worst possible future loss;
- increasing risk after drawdown to accelerate recovery.
A practical ruin-risk review
Define the failure threshold, estimate the strategy's realistic win/loss distribution, include costs and execution uncertainty, model plausible losing sequences, account for simultaneous exposure, and then test how the result changes when the assumptions become worse.
If a small change in assumptions produces a large jump in failure probability, the position size or capital structure is fragile even if the base-case estimate looks acceptable.
Risk of ruin inside the MFXG framework
The parent Risk Management pillar is built around survival before prediction. Risk of ruin turns that principle into a threshold question: can the process remain viable long enough for its edge, if any, to matter?
The goal is not to calculate a reassuringly tiny number. It is to identify whether the combination of edge, variance, leverage and size gives the capital enough room to survive uncertainty.