Risk of Ruin: Why a Winning Strategy Can Still Blow Up an Account

Here is a trader’s nightmare that happens more often than people admit. You have a strategy that wins 55% of the time, your winners are as big as your losers, and on paper you have an edge. Six months later the account is gone. The strategy was fine. The bet size was not. That gap between having an edge and surviving long enough to collect it is what risk of ruin measures: the probability that losses will shrink your capital to a point where you can no longer keep trading.

We will walk through what the term means, where the maths comes from, how bet size changes the answer far more than win rate does, and how an Indian retail trader can use these numbers without needing a quant desk.

What risk of ruin means

Wikipedia’s entry on risk of ruin describes it as a concept from gambling, insurance and finance about the likelihood of losing all of one’s capital, or falling below the minimum needed to keep playing. Investopedia frames it the same way: the probability that losses from investing, trading or gambling become so large that recovery is impossible.

Notice the second half of that definition. Ruin does not have to mean a zero balance. For a trader, ruin is usually the point where the account is too small to take the next position, where margin calls start, or where you simply lose the nerve to follow your system. Wikipedia makes this point too: traders sometimes use the term narrowly to mean losses that push an account below the minimum required to keep making trades.

The simplest example comes straight from the Wikipedia page. Put all your money on one coin toss and your risk of ruin is 50%. One bet, one outcome, half the time you are out.

Where the idea comes from: gambler’s ruin

Risk of ruin grew out of a much older probability puzzle called gambler’s ruin. The earliest known mention is a 1656 letter from Blaise Pascal to Pierre de Fermat, and Christiaan Huygens reworked the problem and published it in 1657 in De ratiociniis in ludo aleae. The set-up: two players start with a fixed stake, money moves from loser to winner after each round, and the game ends when one of them hits zero.

Two results from that work matter to traders:

• A fair game kills the smaller player. A persistent gambler with finite money, playing a game where each bet has an expected value of zero, will eventually go broke against an opponent with unlimited money. The market is that opponent. It never runs out of capital.
• A small edge changes everything, if you size properly. When one player wins each round with probability p and loses with probability q = 1 − p, the gambler’s ruin formula gives each player’s chance of ending penniless. Let the opponent’s stack grow without limit and, for a player with an edge (p greater than q), that formula reduces to (q/p)^N, where N is the number of equal-sized bets your capital can absorb.

That last expression is the one we will use. It assumes even-money bets of a fixed size, which is a simplification, but it shows the shape of the problem better than any long-winded explanation.

An edge tells you which way the account should drift. Bet size decides whether the account is still alive when the drift shows up.

Risk of ruin in numbers: a worked example

Take a hypothetical trader in Mumbai with ₹5 lakh of trading capital. Every trade risks a fixed rupee amount, and every winner pays the same amount the losers cost (a 1:1 reward-to-risk). Their system wins 55% of trades, so p = 0.55 and q = 0.45, and q/p works out to roughly 0.818.

Now change only one thing: how much they risk per trade. N is simply capital divided by risk per trade.

Read that table twice. Same trader, same strategy, same 55% win rate. Risking a fifth of the account per trade gives a better than one-in-three chance of blowing up. Risking 2% makes ruin, in this simple model, a rounding error. Each figure is just 0.818 raised to the power of N: 0.818⁵ is about 0.367, 0.818²⁰ is about 0.018, and so on.

Now shave the edge. Suppose the real win rate is 52%, not 55%, the kind of gap that can open up between a backtest and live trading once brokerage, taxes and slippage bite. Then q/p is 0.48/0.52, about 0.923. At ₹25,000 a trade (N = 20) the risk of ruin jumps to about 20.2%. At ₹10,000 a trade (N = 50) it is about 1.8%. A three-point drop in win rate turned a comfortable 1.8% into one-in-five at the 5% size. Thin edges need small bets.

And at exactly 50%, with 1:1 payoffs, there is no edge at all. That is the fair game from Huygens’ result, and against an opponent with unlimited money the eventual outcome is ruin.

A simpler way to see it: count the losses you can survive

Wikipedia gives an even plainer trading example. Say you have $1,000 you can afford to lose before your broker starts issuing margin calls, each trade can lose at most $200, and each trade has a 50% chance of losing. For four trades or fewer, your risk of ruin is zero, because four losses only use up $800. Over five trades it is about 3%, since all five have to fail: 0.5⁵ is 3.125%. Add more trades and the risk keeps creeping up.

This is the intuition every trader should carry. Your capital divided by your maximum loss per trade is the number of consecutive losers you can take. Make that number small and a perfectly normal bad run finishes you.

How normal are bad runs? With a 45% chance of losing each trade, the chance that any particular stretch of five trades are all losers is 0.45⁵, about 1.8%. That sounds rare, but an active trader taking a few hundred trades a year gets many overlapping five-trade stretches, so a five-loss streak at some point is far more likely than 1.8% suggests. If five losses is all you can absorb, you are betting the account on never having a bad week.

The continuous version: Wikipedia’s approximation

Real trades do not pay exactly +1 or −1. Wikipedia gives an approximate formula for a random walk where each step is drawn from a normal distribution with mean μ and standard deviation σ, starting from a value s:

P(ruin) ≈ ( 2 / (1 + μ/r) − 1 )^s/r, where r = √(μ² + σ²).

Its worked example uses a starting value of 10, a mean of 0.1 and a standard deviation of 1 per step. Then r is about 1.005 and the formula gives a risk of ruin of roughly 13.71%. We reran that arithmetic and got the same answer.

Here is how to translate it for trading. Measure everything in “R”, the amount you risk per trade. Starting capital of 10R means you risk 10% of the account per trade. A mean of 0.1R per trade is a decent positive expectancy. A standard deviation of 1R means individual trade results swing a lot around that average. Even with that real edge, about one trader in seven running this set-up would hit zero before the edge pulled them clear. Double the starting cushion to 20R by halving the bet, and the exponent doubles, so the risk falls to roughly the square of 13.71%, under 2%.

Wikipedia adds a caution worth repeating: the approximation is more accurate when ruin, if it happens, takes many steps. It is not very accurate when one step could make or break you. Which is exactly why you should not size so that one step can.

Why percentage risk changes the picture, and why ruin still matters

Most position-sizing advice says to risk a fixed percentage of current capital, not a fixed rupee amount. Mathematically, that means you never quite reach zero, because each bet shrinks as the account shrinks. Does that make risk of ruin irrelevant? No. It just moves the line from “zero” to “the drawdown I can’t live with”.

Run a 20-trade losing streak through two sizes. At 2% risk per trade, the account keeps 0.98²⁰, about 66.8% of its starting value. At 5% per trade, it keeps 0.95²⁰, about 35.8%. Then add the recovery maths. A 50% loss needs a 100% gain just to get back to where you started. A trader sitting at 35.8% of their capital needs to nearly triple the account to break even, and almost nobody sticks with a system through that. In practice, that is ruin.

Percentage sizing doesn’t stop you hitting bottom. It changes where the bottom is, and you get to choose it in advance.

So define ruin for yourself before you trade. For you it might be a 30% or a 50% drawdown: the point where you would quit the strategy. Then size so that the probability of reaching that line stays small.

What pushes risk of ruin up

• Bet size. The biggest lever by far, as the table shows. Halving the risk per trade doubles N, and because N sits in the exponent, the risk falls much faster than in a straight line.
• Leverage. Wikipedia lists careful control of leverage as one of the main defences. In F&O, a futures lot or a naked options position can risk far more of the account than the margin blocked suggests, so the real N is lower than it looks.
• Unlimited-loss positions. Wikipedia also flags exposure to products with unlimited downside, such as some short positions, where you can lose more than you put in. Short options without a hedge belong here. One gap-up or gap-down open can break the “fixed loss per trade” assumption entirely.
• Correlated positions. Five bank stocks are not five independent bets. If they all fall together on one RBI announcement, you effectively took one large trade. Wikipedia notes that during the 2008 crisis there were stretches when asset classes moved together across regions, which is when diversification helps least.
• An overstated edge. The 55% versus 52% comparison shows how sensitive the result is. Use live or out-of-sample win rates after costs, not your best backtest.

How to keep your risk of ruin low

None of this needs a complicated model. A few rules do most of the work:

1. Cap risk per trade. Pick a small, fixed percentage of capital and stick to it. The table above shows what 2% buys you compared with 10% or 20%. Our guide to the Kelly criterion and position sizing walks through how to turn a measured edge into a bet size.
2. Know your expectancy first. Risk of ruin depends on win rate and payoff together. If you have not measured your average win, average loss and win rate, start with our explainer on trading expectancy. A negative expectancy means ruin is a matter of time, whatever the size.
3. Put the stop in before the trade. The maths assumes a known maximum loss. A stop-loss you move or ignore turns a fixed-risk trade into an open-ended one.
4. Cut size during drawdowns, not after wins. Gambler’s ruin warns that a player who raises the bet after wins but does not cut it after losses will eventually go broke even with a positive edge. Size should follow the account down. The anti-martingale approach is built around this idea, and the opposite habit, doubling after losses, does the reverse.
5. Treat correlated trades as one. Add up the risk on positions that move together and keep the total within your per-trade cap.
6. Set a hard drawdown limit. Decide in advance that at, say, a 20% drawdown you stop, review and trade smaller. That turns a slow bleed into a pause.

For the wider framework around these rules, including stop placement and daily loss limits, see our main guide to risk management in trading.

The limits of the numbers

Be honest about what these formulas leave out. They assume each trade is independent, that wins and losses have a fixed or normally distributed size, and that the win rate stays the same. Markets break all three. Losing trades cluster when conditions change, and real returns have fatter tails than a normal distribution, so extreme days come more often than the model expects. Investopedia notes that black swan events can upend even sophisticated risk models, and that personal risk management is often overlooked or miscalculated. Wikipedia even records that traders disagree on how useful the statistic is, with some calling it close to worthless in practice and others calling it essential.

Our view: treat the number as a floor, not a forecast. If the simple maths says your risk of ruin is 15%, the real figure is probably worse. If it says 0.01%, you have at least ruled out the obvious way to blow up.

The bottom line

Risk of ruin is the probability that losses take you out of the game before your edge has time to work. The single biggest driver is how much you risk per trade. In our hypothetical example, the same 55% strategy went from about a 37% chance of ruin to almost none just by cutting the bet from 20% to 2% of capital. Measure your real expectancy, size small enough to survive a long losing streak, keep correlated bets in check, and decide in advance what “ruin” means for you. A strategy can only make money if the account is still standing.

Disclaimer: This article is for education only and is not investment advice. Trading in stocks and derivatives involves risk, including loss of capital. The worked examples use hypothetical figures. Consult a SEBI-registered adviser before making investment decisions.