Ask most crypto traders what matters most about their system and you’ll hear “win rate.” Ask a quant and you’ll get a different answer: risk of ruin — the probability that a string of losses, at the position size you’re actually using, wipes out your account before your edge has time to play out. A 55% win rate with the wrong position size can still ruin you. A 40% win rate with the right one often won’t. The gap between those two outcomes is a genuinely old piece of mathematics, and it’s worth understanding on its own terms before trusting any calculator’s number.
The Problem Is Three Centuries Old
Risk of ruin isn’t a crypto-era invention — it’s a direct descendant of the gambler’s ruin problem, a classical question in probability theory about a player with finite capital making repeated bets against an opponent with much deeper pockets. Wikipedia’s entry on risk of ruin ties the two directly together, noting that persistent play eventually yields “the stochastic certainty of gambler’s ruin” once you look far enough into the future. The same page gives the closed-form probability for a simple random-walk version of the problem and a concrete illustration: with a $1,000 account, trades capped at a $200 loss, and 50/50 odds, the risk of ruin over five trades works out to roughly 3%.
That example matters because it shows the real lever: it’s not really about whether you’re right more than you’re wrong. It’s about how large a bite a losing streak can take out of your capital before the law of averages gets a chance to work in your favor.
Closed-Form Math Only Goes So Far
The classical formula is clean, but it only holds up under a narrow assumption: fixed bet size, fixed odds, every trade identical. Real trading rarely looks like that — position size often scales with a changing account balance, and win rate and payoff aren’t perfectly constant round to round. Once you leave that narrow case, there’s no tidy formula left to solve, which is exactly the gap simulation is built to fill.
This is where it’s worth looking at how an actual tool handles it, not just the textbook version. Holdy Lab’s Risk of Ruin Calculator is explicit about this trade-off rather than hiding behind a black-box number. Its companion article states the limitation plainly: exact formulas “only exist for simple cases, such as bets of fixed size and fixed odds” — and that a simulation is what “handles risking a share of a changing balance with any win rate and payoff.” The calculator itself describes its method directly: “it plays out thousands of sequences of trades using your win rate, your reward for each unit of risk, the share of the balance you risk per trade, and the number of trades” — specifically 4,000 simulated paths, run on a fixed random seed so a given set of inputs always reproduces the same result rather than a slightly different number every time you check it.
A worked example from the tool: a trader with a 45% win rate, a 1.5 reward-to-risk ratio, risking 2% of the account per trade over 200 trades, lands at roughly a 0.2% chance of hitting a 50%-of-account drawdown. Push that risk to 4% per trade with the same win rate and payoff, and the same source’s companion piece puts the ruin chance at around 8% — a 40x jump in ruin risk from doubling position size, which is a far more dramatic relationship than most traders intuitively expect.
It’s also honest about where this still simplifies reality: the method still assumes “every trade is independent and has the same odds” — which is a real limitation (strategies with momentum, mean-reversion, or regime-dependent edge violate it to varying degrees), stated by the tool itself rather than left for the user to discover the hard way.
The Sizing Question This Connects To
Once you know your risk of ruin at a given position size, the natural next question is what size you should be using — and that has its own, older piece of math behind it: the Kelly criterion. Formulated by John L. Kelly Jr. in a 1956 Bell Labs paper and only later adopted by gamblers and investors, the formula in its basic bet-sizing form is f = p/l − q/g, where p is your win probability, q is your loss probability (1−p), g is what you gain per unit on a win, and l is what you lose per unit on a loss. Holdy Lab’s own Kelly Criterion Calculator applies this directly: at a 45% win rate and a 2:1 payoff, full Kelly comes out to roughly 17.5% of the account per trade — a number most working traders would consider aggressive, which is exactly why half-Kelly (about 8.75%) and quarter-Kelly (about 4.4%) are commonly used instead, trading some growth rate for a meaningfully smoother equity curve.
Used together, the two tools answer different halves of the same question: Kelly tells you the size that maximizes long-run growth under your stated edge; risk of ruin tells you how survivable that size (or any other size you actually plan to use) is before that long run arrives.
Why This Is Worth Taking Seriously
None of this requires trusting a single number blindly. The value is in what it exposes: that position size, not win rate, is usually the variable doing the most damage to an account, and that the relationship between size and ruin risk isn’t linear — it accelerates. A system can have genuine positive expectancy and still be unsurvivable at the size someone is actually trading it, and risk-of-ruin math is the tool that makes that visible before it happens rather than after.
Frequently Asked Questions
Is risk of ruin the same thing as max drawdown?
No — max drawdown describes what actually happened in a specific historical or backtested run (the worst peak-to-valley decline observed). Risk of ruin is forward-looking and probabilistic: given your stated win rate, payoff, and position size, what’s the chance of hitting a catastrophic loss level at all, across many possible futures, not just the one that already happened.
Why does a small change in position size cause such a large change in ruin risk?
Because losses compound against a shrinking base — a 2% loss on a smaller balance is a smaller dollar amount, but you need a larger percentage gain to recover from a given percentage loss, and that asymmetry gets sharply worse as losses stack. Simulation captures this compounding effect directly; a simple formula that assumes fixed bet size doesn’t.
Does using Monte Carlo simulation make the result exact?
No, and no honest implementation claims it does — it’s an estimate built from thousands of simulated paths, bounded by whatever assumptions go into the simulation (independence between trades being the main one). What it buys you over a closed-form formula is the ability to model a changing balance and non-fixed odds at all, not false precision.
Should a beginner use full Kelly sizing?
Most practitioners don’t, and the math itself explains why: full Kelly maximizes long-run geometric growth, but at the cost of large, psychologically brutal swings along the way — which is precisely why fractional Kelly (half or quarter) is the more commonly recommended starting point in practice, trading some theoretical growth rate for a drawdown profile an actual human can sit through without abandoning the system.