Poker Variance Calculator

Variance is the natural swing of results around your true winrate. Even a clear winner endures long break-even or losing stretches; this calculator simulates 1,000 alternate versions of the exact sample you enter, so you see the realistic spread of outcomes instead of guessing at it.

Far more than most players expect. With a standard deviation around 120-150 bb/100, a solid 5 bb/100 PLO winner still shows a loss over a 50,000-hand stretch about one time in five, and downswings of 15-25 buy-ins are routine. Run your own numbers above and look at the shaded band: 95% of runs is a much wider corridor than intuition suggests.

Size it from risk of ruin, not a flat buy-in rule. At 3 bb/100 with SD 140, keeping risk of ruin under 5% across your next 100,000 hands takes about 65 buy-ins, and the classic play-forever formula says nearer 100. The 20-30 buy-in habit imported from NLHE is far too thin for PLO. The requirement also moves fast when your winrate estimate is off: half the winrate, double the roll. The bankroll card above computes all of this for your own numbers.

Risk of ruin is the probability your bankroll ever hits zero, given your winrate, standard deviation and starting roll. The calculator counts it directly from the simulated trajectories (the lowest point each run reaches) and labels the horizon it covers, rather than leaning on a single closed-form approximation. Target under 5%; even 5% is on the high side for a pro, and the standard discipline is to move down once a third to half of the roll is gone so you stay under it.

Take it from your tracker (PokerTracker or Hold'em Manager): the per-100-hands standard deviation, ideally the all-in-adjusted figure. As a rough guide it runs around 80-100 bb/100 for NLHE 6-max and 120-150+ for PLO. The presets here default to 140 for PLO 6-max.

Deeper than feels plausible. At 3 bb/100 with SD 140, about nine runs in ten hit a 25 buy-in drawdown somewhere in 100,000 hands, and break-even stretches of tens of thousands of hands are normal. Depth scales with your standard deviation and shrinks, relative to expectation, as your edge grows. The Downswings card reports the worst trough of every simulated run, so you can see what is normal for your numbers before reading anything into your own graph.

Much higher. In a standard structure you cash under a fifth of the time and rare deep runs carry most of the profit, so long losing stretches are the default experience even for good players. Switch the calculator to tournament mode to model it from your ROI, field size and payout structure instead of a cash winrate.

Yes, and over normal samples it is common. Results across a few tens of thousands of hands are dominated by variance, not skill, so a true winner can show a loss and a true loser a profit. The Bayesian 'Am I a winner?' card turns the question around: given your observed winrate and sample size, it estimates the probability you are actually a winner, which is far more honest than eyeballing a graph.

Being staked moves the downside to your backer: you play on their money, carry makeup on losses, and split profits once the makeup clears. Your own money stops being at risk, and in exchange your expected take shrinks by the backer's cut. The backing card simulates exactly that deal structure, including how often an engagement ends still in makeup and the backer's own risk of ruin.

No. Your standard deviation per 100 hands stays roughly the same when you add tables; what changes is your winrate, which drops as attention splits. Same variance on a smaller edge means more frequent and deeper downswings, even when total hourly EV holds. More hands per hour does compress the long run in calendar time, so the trade can still be right while the winrate holds up. The multi-tabling penalty in settings models it the honest way: winrate down, SD unchanged.

Four cards per player means ranges run much closer in equity. Made hands are rarely locks before the river, draws are huge, and the money regularly goes in with both sides live, so the same edge produces far bigger swings around it. That extra standard deviation per 100 hands is exactly what this calculator turns into downswing, risk-of-ruin and bankroll numbers.

Use your real sample, or the volume you plan to play. The point is to see how wide the outcome band is at that scale: at 10,000 PLO hands the spread is enormous, by 500,000 it narrows a lot but never vanishes. The 'Am I a winner?' card will also tell you how many hands pin your winrate to within 1 bb/100 at 95% confidence: with PLO variance, about 7.5 million.