Risk of ruin is the probability that a trading account falls so far that its owner can no longer keep trading. It is one of the few numbers in trading that describes survival instead of profit, and it depends on position size far more than most traders expect. A strategy with a real edge can still carry a high risk of ruin if every trade puts too much of the account on the line.
This guide covers what the term means, how the classic formula works and where it stops being useful, how to estimate the figure with a simulation, and which decisions bring it down.
What Risk of Ruin Actually Measures
In its textbook form, risk of ruin is the chance of losing the entire stake. Traders use a broader and more practical definition: the chance that the account reaches a loss from which recovery is unrealistic, or at which trading simply has to stop. That point arrives well before zero. It may be a personal limit, a margin call from the broker, or the maximum loss written into the rules of a funded account.
The usual yardstick is the drawdown, the decline from a previous peak in the account's value to a later low. Drawdowns are harder to repair than they look, because the gain needed to recover is always larger than the loss that came before it:
- a 10% loss needs an 11.1% gain to get back to the starting point;
- a 20% loss needs 25%;
- a 30% loss needs 42.9%;
- a 50% loss needs 100%;
- a 75% loss needs 300%.
The deeper the hole, the harder the arithmetic works against the trader, and the stronger the temptation to raise the stakes to climb out. Risk of ruin puts a number on the likelihood of getting there in the first place, and calculating it forces three inputs into the open: the percentage of winning trades, the size of the average win relative to the average loss, and the share of capital risked on each trade. The first two describe the strategy, and together they give its expectancy per trade. The third is a decision, and it turns out to matter most.
The Classic Formula and What It Assumes
The problem is nearly four centuries old, and it was not invented for markets at all. Blaise Pascal posed it in a letter to Pierre de Fermat in 1656, and Christiaan Huygens published a version the following year in the first printed book devoted to probability. It is known as the gambler's ruin. The version behind the trading formula is the lopsided one: a player with a fixed stake makes a series of equal bets against a much richer opponent, and the question is how likely the player is to lose everything.
The form that circulates in trading books and on trading websites is:
Risk of ruin = ((1 − A) / (1 + A)) ^ C
A is the trader's edge, meaning the probability of a win minus the probability of a loss. C is the number of capital units, meaning how many full-size losses the account can absorb. With a win probability p and a loss probability q, the fraction in brackets is simply q divided by p, which is the textbook gambler's ruin result in another notation. The formula assumes independent trades, wins and losses of the same size, the same fixed amount at risk every time, and ruin defined as losing the whole stake.
Take a strategy that wins 55% of the time with equal wins and losses, so that A is 0.10. Risking a tenth of the starting capital on each trade gives ten units, and the formula returns a risk of ruin of 13.4%. Risking a twentieth cuts it to 1.8%. Risking a fiftieth, or 2% of the starting capital, brings it down to 0.004%. The edge never changed. Only the bet size did, and the probability of ruin fell by a factor of more than three thousand.
Two cautions apply. The first concerns a mistake that survives on widely read trading pages, which define the edge as the win rate itself. Plugging 0.55 into A instead of 0.10 turns the 13.4% of the example into 0.0004%: a thirty-thousandfold error, all of it in the comforting direction. The second is that a thin edge offers little protection: with a 52% win rate, ten units leave a 44.9% risk of ruin, and even fifty units leave 1.8%.
The limits of the formula are just as clear. Real trades do not come with equal wins and losses, position size is commonly set as a percentage of the current balance instead of a fixed amount, and nobody waits for zero before stopping. Closed-form formulas exist beyond the coin-flip case, including one based on the work of Cox and Miller that needs only the mean and standard deviation of the results. Each of them buys its tidy answer with assumptions: a fixed position size, independent trades and a stable distribution of outcomes. They are useful as a first estimate, and they cannot replace an analysis of the actual strategy.
Estimating Risk of Ruin With a Simulation
For a real strategy, the practical route is simulation. The starting point is the strategy's own record: win rate, average win, average loss and risk per trade, taken from a sample of trades large enough to mean something. Two steps follow.
The first is to define ruin before running anything. A drawdown of 30% or 40% is a common choice, but the threshold is personal. For some traders 20% is already unacceptable, while for an aggressive strategy it may be an expected phase. The horizon has to be fixed as well, because over an unlimited number of trades every drawdown eventually happens.
The second is to generate many possible futures. A Monte Carlo test does this by building thousands of alternative sequences from the historical trades. Reshuffling the same trades into a new order leaves the final result unchanged and shows how much the drawdown depends on sequence alone. Resampling with replacement, where a trade can be drawn more than once, varies the final result as well. Either way the logic is the same: an edge says nothing about the order in which wins and losses arrive, and a cluster of losses does far more damage than the same losses spread over a year. The share of simulated paths that touch the threshold is the estimated risk of ruin.
A worked example shows what the exercise reveals. Consider a system that wins half of its trades, with an average win 1.5 times the average loss. Its expectancy is a healthy 0.25 times the amount risked per trade. Each trade risks a fixed percentage of the current balance, and ruin is defined as a 30% drawdown from the account's peak at any point in the next 250 trades, about a year at one trade per business day. Simulating 200,000 paths for each position size gives the following probabilities:
- risking 1% per trade: about 0.01%;
- risking 2%: about 4%;
- risking 3%: about 30%;
- risking 5%: about 91%;
- risking 10%: a practical certainty.
Nothing about the strategy changes from one line to the next. The same profitable system is safe at 1%, uncomfortable at 3% and almost sure to suffer a 30% drawdown within a year at 5%. Extending the horizon to 1,000 trades raises the 2% figure to about 19% and the 3% figure to about 79%, which is why a risk of ruin always has to be read together with its time frame.
A weaker edge shifts everything. Lower the win rate from 50% to 45%, which still leaves the system profitable on paper, and the probability of a 30% drawdown within 250 trades at 2% risk rises from about 4% to about 25%.
The Inputs Are Estimates Too
A calculation of this kind is only as good as the statistics fed into it, and those come from a finite sample. A win rate of 60% measured on 20 trades is compatible, at the usual 95% confidence level, with a true win rate anywhere between about 39% and 78% (Wilson score interval). On 100 trades the range narrows to about 50% to 69%, and on 300 trades to about 54% to 65%. A test based on twenty trades says very little, and it tends to produce false confidence.
Live trading adds frictions that a backtest usually understates: spreads and commissions, slippage (the gap between the expected price and the executed one), shifts in volatility, operational mistakes, and market phases unlike anything in the test period. On balance they push the real risk of ruin above the calculated one. The sensible response is to combine tools: strategy statistics, simulations, drawdown analysis and control of position size. A perfect number is out of reach and would only offer false comfort. What the exercise can show is whether the system is robust or whether one bad sequence is enough to put the account in trouble.
Losing Streaks Are a Matter of Arithmetic
Most traders underestimate how long a losing streak a sound strategy will produce. With a 50% win rate, the probability of seeing at least seven consecutive losses somewhere in 250 trades is about 63%. For eight in a row it is about 38%, and for ten in a row about 11%. Even with a 60% win rate, a run of seven losses turns up in about 22% of 250-trade samples.
Position size decides what such a streak costs. Seven straight losses take 6.8% off an account that risks 1% of the current balance per trade, 13.2% at 2%, 30.2% at 5% and 52.2% at 10%. At 5% per trade, an event that is more likely than not within 250 trades is enough, on its own, to produce a 30% drawdown.
How to Lower Your Risk of Ruin
Reducing risk of ruin does not mean eliminating risk, which is impossible in trading. It means building conditions that let the account survive its bad phases without rewriting the plan every week or raising the stakes to win losses back, the logic that makes the martingale so destructive. A few habits do most of the work.
- Risk a small, fixed fraction on every trade. Risking 5% or 10% of the account can look acceptable while the strategy is working and becomes dangerous during a run of stop-losses. A range of 0.5% to 2% per trade slows the speed at which an account can deteriorate, as the simulation above shows. The position size that matches a chosen percentage depends on the stop distance and the pip value, and a lot size calculator handles the conversion.
- Judge the strategy by more than its win rate. A strategy that wins 70% of its trades can still be fragile if the average loss is much larger than the average win. A strategy with a lower win rate can work if the winners more than pay for the losers. Risk of ruin depends on the combination of win probability, average win and average loss.
- Set a maximum drawdown in advance. A drawdown limit defines when to stop, cut exposure or review the system. Without one, the usual outcome is to keep trading while the account is already sending clear signals. The limit has to fit the strategy: too tight and it halts the system at the first normal losing phase, too loose and it triggers when the damage is already hard to repair.
- Test across different market conditions. A system tried only on a favorable phase looks stronger than it is. Testing on periods with different volatility, trends, ranges and stress gives a more honest picture. A demo account can be enough for this, provided it is not used only to confirm conclusions already reached.
- Treat the growth-optimal size as a ceiling. The Kelly criterion gives the fraction of capital that maximizes long-run growth, and for the example system above it comes out at 16.7% of the account per trade. Edward Thorp, the mathematician who applied the formula to blackjack and then to the stock market, worked out what that costs: at full Kelly, the chance of seeing the account cut in half at some point is one in two. At half that size it drops to one in eight, in exchange for giving up a quarter of the growth rate. The size that is optimal on paper sits far above what most traders can live through, and because the inputs are estimates, erring on the small side is the safer mistake.
The same arithmetic governs funded-account evaluations, where ruin is written into the rules. One widely used evaluation sets a profit target of 10% in its first stage and 5% in a second, and closes the account in either stage at a total loss of 10% of the initial balance. For the example system, risking 1% of the initial balance per trade clears the first stage about 97% of the time and both stages about 94%. Risking 2% lowers the odds of clearing both stages to about 74%, and risking 5% to about 46%. These figures ignore costs and the daily loss limit that such programs add, so the real odds are lower. A system with no edge and equal wins and losses clears the first stage half the time and both stages about a third of the time, whatever the position size, and published prop firm pass rates sit far below even that.
Survival Comes First
The question behind risk of ruin is simple: how likely is it that this way of trading ends the account before the edge has time to show? An exact answer is out of reach, because the inputs are estimates and markets change. An honest approximate answer is available to anyone with a trade history and the patience to resample it, and the most reliable way to improve it is also the least glamorous one: smaller positions.
