A winrate on its own tells you almost nothing about what a single sample of poker actually looks like. Two players with identical long-run edges can have wildly different results over 100,000 hands, purely because of variance.
Loading calculator...
The Poker Variance Calculator converts your winrate and standard deviation into a real confidence range, so you can see how likely a losing stretch is even when your underlying game is profitable.
It covers both cash games, using the standard bb/100 framework, and tournaments, using buy-ins as the unit, since the two formats behave very differently over a given sample size.
π How to Use the Poker Variance Calculator
Select Cash Game or Tournament mode first, since the inputs and units are different for each format. Cash game variance is measured in big blinds per 100 hands; tournament variance is measured in buy-ins.
Standard deviation, not winrate, is what determines how wide your realistic range of outcomes is β two players with the same winrate can have very different variance depending on game type and stakes.
Enter your best estimate of your own standard deviation if you track your results, or use a typical figure for your game type as a starting point, then choose a confidence level to see your likely range.
π’ Calculator Fields Explained
Winrate (bb/100) – your average profit per 100 hands, expressed in big blinds, for cash game mode.
Standard Deviation (bb/100) – a measure of how much your results swing around your winrate; typical full-ring is 60-80, six-max is 80-100+.
Number of Hands – the size of the sample you want to project variance over.
Big Blind Size – your stake’s big blind in currency terms, used to convert bb results into cash.
ROI (%) – your average tournament return on investment as a percentage of buy-in, for tournament mode.
Average Buy-In – the typical entry cost of the tournaments you play.
Standard Deviation (buy-ins) – how much a single tournament result swings, typically 50-100+ buy-ins depending on field size and structure.
Number of Tournaments – the sample size you want to project variance over.
Confidence Level – how wide a range you want (higher confidence produces a wider range).
π° Understanding the Results
| Result Field | What It Tells You |
|---|---|
| Expected Profit | Your average projected result over this sample, based on your stated winrate/ROI |
| Standard Deviation (Sample) | How much your actual result could reasonably swing from the expected profit |
| Confidence Range | The band your result is statistically likely to fall within, at your chosen confidence level |
| Probability of a Losing Sample | The estimated chance you end this exact sample size in the red despite a positive winrate |
The confidence range is the single most useful number here, because it reframes “winrate” as a distribution of plausible outcomes rather than a guaranteed result.
A positive expected profit does not mean a losing sample is impossible β over a small enough sample, even a strong long-term winrate can still show a real chance of a net loss.
Probability of a Losing Sample is the number that best explains why a genuinely winning player can still go through a demoralizing multi-month downswing purely from normal variance.
π Calculation Formulas
| Metric | Formula | Unit |
|---|---|---|
| Expected Profit (Cash) | winrate Γ (hands Γ· 100) | bb |
| Sample Standard Deviation | stdDev Γ β(hands Γ· 100) | bb |
| Confidence Interval | mean Β± (z-score Γ sample stdDev) | bb or buy-ins |
| Expected Profit (Tournament) | ROI% Γ number of tournaments | buy-ins |
Standard deviation grows with the square root of the sample size, not linearly, which is why doubling your hands played only widens your dollar-variance by about 41%, not 100%.
This square-root relationship is also why larger samples make your winrate estimate more reliable relative to its size, even though the absolute dollar swings keep growing.
π Practical Examples
Example 1 – Six-Max Grinder: 5 bb/100 winrate, 90 bb/100 stdev, 100,000 hands, $1 big blind, 95% confidence. Expected profit = 5,000 bb ($5,000). Sample stdev β 90 Γ β1000 β 2,846 bb.
Example 2 – Same Player, Smaller Sample: Same rates over just 10,000 hands. Expected profit = 500 bb ($500), but sample stdev β 90 Γ β100 = 900 bb β larger than the expected profit itself.
This is exactly why 10,000 hands is considered far too small a sample to draw firm conclusions about a winrate β the noise is larger than the signal at that size.
Example 3 – MTT Grinder: 15% ROI, $50 average buy-in, 75 buy-in stdev, 1,000 tournaments played. Expected profit = 150 buy-ins ($7,500), sample stdev β 75 Γ β1000 β 2,372 buy-ins ($118,600).
Example 4 – Same MTT Grinder, One Year In: Same rates but only 200 tournaments played. Expected profit = 30 buy-ins ($1,500), but the 95% confidence range can easily span from a large loss to several times the expected profit, showing why MTT results need very large samples to trust.
π‘ Tips & Best Practices
Track your own actual standard deviation from your hand history database rather than relying on generic averages, since it varies a lot by stakes, game format, and playing style.
Treat any sample under roughly 50,000-100,000 hands as noisy for cash games, and under a few thousand tournaments as noisy for MTTs, when judging your true winrate.
Use a wider confidence level like 95% or 99% when making bankroll decisions, since you want to plan for realistic worst cases, not just the average scenario.
Running this calculator before a big downswing hits can make the eventual downswing far less psychologically damaging, since you already know it falls inside your expected range.
Higher-variance formats like MTTs and short-handed cash games need proportionally larger bankrolls and longer sample sizes before conclusions about skill become reliable.
Re-run the calculator periodically as your sample grows, since your estimated standard deviation itself becomes more accurate with more data.
- Separate your stats by game type and stake, since mixing formats distorts your true standard deviation
- Recalculate after major downswings or upswings to recheck whether they fell inside your expected range
β οΈ Common Mistakes to Avoid
Judging Winrate Off a Tiny Sample
Players frequently draw firm conclusions about their skill level from a few thousand hands or a few dozen tournaments.
Treating a small sample’s results as proof of your true winrate, rather than as one draw from a wide distribution, is one of the most expensive misunderstandings in poker.
Use the confidence range from this calculator to see just how wide the plausible outcomes are at your actual sample size.
Ignoring Standard Deviation Differences Between Formats
A stdev appropriate for full-ring cash games is badly wrong if applied to six-max or to tournaments, which swing far more per unit played.
Using a cash-game-typical standard deviation to model tournament variance will drastically understate your realistic downswing risk.
Always match your standard deviation input to the actual format and stakes you’re modeling.
Confusing Expected Profit With Guaranteed Profit
Expected profit is a long-run average, not a promise about any specific sample.
Mistaking expected profit for a guarantee is what leads players to over-leverage their bankroll right before a statistically ordinary downswing that was well within their calculated range all along.
Not Recalculating After Big Sample Changes
An estimate of standard deviation from a small sample can itself be unreliable, but many players never revisit it once they’ve played more.
Update your inputs periodically so the confidence range reflects your actual current data, not an outdated early-career estimate.
π― When to Use This Calculator
Use this calculator whenever you want to understand whether a recent downswing (or upswing) falls within normal variance, or when planning bankroll requirements for a given format.
Experienced players often say that understanding variance is what separates those who survive a full career in poker from those who quit during an ordinary, statistically expected downswing.
It’s also useful before increasing stakes, since it shows how much wider your realistic swings will become at a higher big blind size or buy-in.
π Related Calculators
Poker Rake Calculator, Rakeback Calculator, ROI Calculator, ICM Calculator, Bankroll Management Calculator
π Glossary
Variance – the statistical measure of how spread out poker results are around the expected value.
Standard Deviation – the square root of variance, expressed in bb/100 for cash games or buy-ins for tournaments.
bb/100 – big blinds won or lost per 100 hands, the standard cash-game winrate unit.
Downswing – a losing stretch that falls within the normal range of variance for a given winrate and sample size.
Confidence Interval – a statistically derived range within which a result is likely to fall at a given probability level.
Z-Score – a value from the normal distribution used to construct a confidence interval at a chosen confidence level.
ROI – Return on Investment, a tournament player’s average profit as a percentage of total buy-ins.
Sample Size – the number of hands or tournaments used to estimate a winrate or ROI.
Six-Max – a cash game format with six players per table, generally higher variance than full ring.
Full Ring – a cash game format with up to nine or ten players, generally lower variance than six-max.
Normal Distribution – the bell-curve statistical model used to approximate the distribution of poker results.
Bankroll Management – the practice of sizing your playing bankroll to survive expected variance without going broke.
β Frequently Asked Questions
How many hands do I need before my winrate is reliable?
Most serious players consider 100,000+ hands a reasonable minimum for cash games, with some formats needing even more depending on stakes and stdev.
At 100,000 hands with a typical six-max stdev of 90, the confidence range can still span several thousand dollars either side of your expected profit.
Why is tournament variance so much higher than cash game variance?
Tournaments concentrate almost all of your profit into a small number of deep finishes, so the vast majority of events lose money by design.
A player with a genuinely strong 20% ROI can still go hundreds of tournaments without a major cash, purely from the top-heavy prize structure.
This is why tournament samples need to be far larger, often thousands of events, before ROI becomes statistically trustworthy.
Can this calculator tell me if I’m actually a winning player?
It can’t prove your true long-run winrate, but it can show you how much your current sample’s result could plausibly differ from that true rate.
If your actual results sit well outside the calculated confidence range for your assumed winrate, that’s a signal your true winrate may be different from what you assumed.
Does a higher confidence level always give a more useful answer?
A higher confidence level widens the range, which is more conservative but also less precise about the most likely outcome.
Many players use 95% for general planning and reserve 99% specifically for worst-case bankroll stress-testing.
Should I use my own tracked stats or a generic standard deviation figure?
Your own tracked standard deviation, from a large enough sample, is always more accurate than a generic figure for your specific style and stakes.
A loose-aggressive six-max player might run a stdev well above 100 bb/100, while a tight full-ring grinder might sit closer to 60, so generic figures are only a rough starting point.
βοΈ Legal Disclaimer
This calculator is provided for informational and educational purposes only. It uses statistical approximations and does not constitute financial or gambling advice or a guarantee of future results. Please gamble responsibly.
