What it is

Risk of ruin is the probability that a sequence of losses will reduce your capital so far that you can no longer trade - either because the account is wiped out, or because it falls below the minimum needed to keep going, or because the drawdown is so deep you stop. It is the most important risk in trading, because it is the one risk you cannot recover from. Every other mistake can be corrected next week; ruin ends the game.

The Kelly criterion is a formula, born in information theory and gambling, that answers a precise question: given an edge, what fraction of your capital should you stake on each opportunity to maximise the long-run growth rate of that capital? It is the mathematical bridge between having an edge and sizing trades to exploit it without destroying yourself. Together, risk of ruin and Kelly are the two faces of position sizing: ruin is the danger, Kelly is the discipline that bounds it.

How it works

Risk of ruin depends on three things, and understanding their interaction is the whole lesson.

• Edge. A positive expectancy pushes risk of ruin down; a negative one drives it inexorably toward certainty. With no edge and enough bets, ruin is guaranteed - this is why a casino always wins eventually against a gambler.
• Bet size relative to capital. This is the lever you control. Risking 1% of capital per trade and risking 25% per trade can share the identical edge yet have wildly different ruin probabilities. Large bets make ruin likely even with a real edge, because a normal losing streak - which a real edge still produces - can compound into catastrophe.
• Number of trades. Over many trades, a positive edge tends to assert itself, but only if you survive long enough to get there. Size too large and a losing streak removes you from the game before the edge can pay off.

The interaction of these three is where intuition fails most people. The killer fact is that losses and gains are not symmetric in their effect on capital. A 50% loss requires a 100% gain just to break even; a 75% loss requires a 300% gain. This asymmetry means large bets are doubly punishing: not only is ruin more likely, but even the drawdowns short of ruin are disproportionately hard to climb out of. Two traders with the identical edge - say a coin that pays slightly more than even money - can end up at opposite fates purely through bet size: the one risking 2% per flip compounds steadily upward, while the one risking 40% per flip almost certainly blows up, because a perfectly normal run of bad flips, which the edge does nothing to prevent, compounds into an unrecoverable hole. Edge tells you whether to play; bet size tells you whether you survive long enough for the edge to matter.

The Kelly intuition follows directly. Kelly says: bet more when your edge is larger and your odds are better, and bet less when the edge is thin. For a simple bet, the Kelly fraction is roughly edge divided by odds - the bigger your statistical advantage, the larger the optimal stake; the smaller the advantage, the smaller the stake. Kelly is the size that maximises the growth rate of capital over the long run. Bet smaller than Kelly and you grow more slowly but more safely; bet larger than Kelly and, counterintuitively, you grow slower and more dangerously, because oversized bets amplify drawdowns faster than they amplify gains. Full Kelly is the knife-edge beyond which more aggression actively hurts you.

How to use it

1. Confirm you actually have an edge. Kelly applied to a negative-expectancy strategy tells you to bet zero - and any positive size accelerates ruin. Sizing is downstream of having a genuine, measured edge.
2. Estimate the Kelly fraction from your win rate and payoff ratio, but treat it as a ceiling, not a target.
3. Bet a fraction of Kelly - typically a half or a quarter. This is the standard professional practice, for reasons below.
4. Recompute as your edge estimate changes, and shrink size when uncertainty about the edge is high.

Worked example. Suppose a setup wins 55% of the time and pays 1:1 (winners and losers the same size). The simplified Kelly fraction is roughly the edge: bet about 10% of capital per trade. But that 10% assumes your 55% win rate is exactly right and stable forever. If reality is 52%, full Kelly on a wrong estimate is wildly oversized, and the resulting drawdowns are brutal - full Kelly routinely produces 50%+ peak-to-trough declines even when the edge is real. Halving it to 5% (half-Kelly) cuts expected growth only modestly while dramatically reducing the depth of drawdowns and the risk of ruin. That trade - give up a little growth to buy a lot of safety - is why professionals almost never bet full Kelly.

There is a subtle but important gap between the textbook Kelly formula and how a stock trader actually sizes a position. The pure formula assumes a fixed bet with known odds, like a coin or a roulette wheel. A real trade has no such clean odds - the "win" and "loss" amounts depend on where you place the stop and target, the win rate is estimated and uncertain, and outcomes are not the two tidy values the formula expects. So in practice traders rarely plug numbers into the literal Kelly equation; instead they use its spirit. The spirit says: risk a small, fixed fraction of capital per trade (the 0.5%-1% rule you have met is a deeply fractional-Kelly choice), scale that fraction up only when your edge is clearly stronger, and never let the fear of missing growth push you toward the full-Kelly cliff. Kelly is best understood not as a calculator you run before every trade, but as the proof that small, consistent fractional risk is mathematically superior to the aggressive sizing most beginners are tempted toward.

Strengths & limits

The strength of the Kelly framework is that it makes the link between edge and size rigorous. It proves that there is an optimal bet size, that betting more than it is strictly worse on both growth and risk, and that the right size scales with the strength of your edge. It reframes position sizing from a vague comfort question into a quantitative one. Perhaps its single most valuable lesson is counterintuitive and worth stating plainly: past the optimum, more aggression lowers your returns. Most beginners assume that betting bigger, on a genuine edge, simply means winning faster with more risk. Kelly proves that beyond a point it means winning slower and with more risk - the worst of both worlds. That alone reorients how a serious trader thinks about size.

The limits are why fractional Kelly dominates real practice. Full Kelly is optimal only if your edge estimates are perfectly accurate and stationary - and they never are. Win rates are measured on finite samples and drift over time; a slightly overestimated edge makes full Kelly catastrophically too large. Full Kelly also tolerates drawdowns most humans cannot psychologically survive: the mathematically optimal path can include a temporary loss of half your capital, and few traders hold their nerve through that. Fractional Kelly - betting a half, a third, or a quarter of the formula's output - sacrifices a small, well-understood amount of theoretical growth in exchange for far shallower drawdowns, a much lower risk of ruin, and robustness to the inevitable errors in your edge estimate. In a world where you never truly know your edge, betting smaller than optimal is the optimal response to that uncertainty. Put differently: full Kelly is the right answer to a question - "how much should I bet if my edge is known exactly?" - that no real trader can ever truthfully answer yes to. Fractional Kelly is the right answer to the question you actually face, which is how to bet when your edge is only estimated, possibly decaying, and certainly noisier than your spreadsheet suggests.

Key takeaway: Risk of ruin is the chance of losing enough capital to be forced out, driven by edge, bet size, and number of trades; Kelly gives the growth-maximising bet size for a known edge - but because real edges are uncertain and full Kelly's drawdowns are brutal, professionals bet a fraction of Kelly to trade safety for only a little growth.