Poker Variance Calculator
Downswings, risk of ruin, and the honest range of your results — for live hours or online hands.
The band is the math (5th-95th percentile of outcomes); the gray lines are 3 simulated runs at exactly your win rate — this is what running good and running bad look like.
Probability of a peak-to-trough downswing at least this deep somewhere in your 100,000 hands (3,000 simulated runs, deterministic).
Expected profit, the outcome bands, loss probability, risk of ruin, and bankroll requirements are exact formulas over the standard model of a win rate (Brownian motion with drift — the same model every serious variance calculator uses). Worst-downswing odds have no simple formula, so they come from 3,000 simulated runs with a fixed seed: the same inputs always show the same numbers.
What is variance in poker?
Variance is the gap between your true win rate and what any finite stretch of play actually pays you. A solid winner with a real edge still books losing weeks, losing months, and downswings deep enough to feel like proof the edge is gone. The point of this page is to put numbers on that: given your win rate and volatility, how wide is the range of legitimate outcomes, how deep do normal downswings run, and how much bankroll makes going broke a non-issue.
The output that surprises people most is the win-rate certainty tile. Confidence in a measured win rate narrows with the square root of the sample, so even 100,000 hands leaves an honest uncertainty of several big blinds per 100 — a winning graph and a break-even graph can be the same player. For what live players should track instead of a hand-picked bb/100, see the honest win-rate article.
Where do I get my standard deviation?
Online, your tracker reports it directly; around 100 bb/100 is typical for 6-max no-limit cash. Live, compute it from your own session log: the standard deviation of your hourly results is the $/√hr figure this calculator wants. If you track sessions in PokerInk, your logged results already contain everything needed — the more sessions, the better the estimate. When in doubt, run the calculator twice with a low and a high guess; the qualitative picture rarely changes.
How much bankroll do I need?
The bankroll tile answers it for your exact numbers: the formula σ²·ln(20)/(2·win rate) caps the lifetime risk of ruin at 5%. Two things follow from the shape of that formula. Halving your win rate doubles the bankroll you need, and moving to a wilder game grows it with the square of the volatility — which is why bankroll rules of thumb quoted in buy-ins differ so much between games. A losing or break-even player has no bankroll large enough; the formula only exists for winners.
Is a big downswing evidence I'm playing badly?
Check the downswing table before concluding anything. If a downswing of your size shows a double-digit probability over your sample, the cards alone explain it and the correct response is bankroll discipline, not a rebuild of your game. If its probability is well under a few percent, treat it as a prompt to review hands — variance is a possible explanation, just no longer the likely one. Reviewing beats guessing either way: check the big all-ins from the downswing and let the equities say whether you kept getting it in good.
Scenarios are shareable: your inputs are written into the page URL, so copying the address bar links straight to this exact scenario. The model assumes results are independent with a stable win rate — game selection, tilt, and moving stakes are outside it.