01What variance actually means in poker

Variance is the reason a winning player can lose for weeks and a losing player can run hot for months. Skill sets the direction of a graph, but luck sets how far the line wanders from that direction at any given moment. The variance calculator makes that wandering measurable. Instead of a single profit figure it produces a range, and the width of that range is often the most revealing number on the screen.

The two inputs that drive everything are win rate and standard deviation, both expressed per 100 hands. Win rate is the edge - how many big blinds the player expects to gain on average. Standard deviation is the noise - how far a typical 100-hand block strays from that average. A win rate of five big blinds against a standard deviation of ninety tells the whole story at a glance: the edge is real but small, and the noise around it is enormous.

In plain terms, the calculator shows how far real results can drift above or below your expected win rate even when your edge is completely genuine. The same skill can produce a smooth climb, a flat year, or a long dip into the red, purely depending on how the cards happen to fall over the sample you are looking at.

02The formulas behind the readout

Expected profit scales linearly with volume, while the spread scales with the square root of volume. That single difference is why edges win in the long run and why the short run feels so random.

Expected profit (bb) = win rate x hands / 100. Standard deviation over the sample = SD per 100 x square root of (hands / 100). The 70 percent interval is the expected value plus or minus about 1.04 standard deviations; the 95 percent interval is plus or minus 1.96.

Because profit grows with hands but the spread grows only with the square root of hands, doubling the sample roughly doubles the expected gain while increasing the swing by only about forty percent. Over a large enough sample the edge overwhelms the noise. Over a small one, the noise wins, which is the mathematical core of why poker punishes impatience.

03How to read the confidence band

The chart draws the expected line climbing from zero to the projected profit, with the shaded 95 percent band widening around it. The key insight is the lower edge of that band. For a modest win rate over a realistic sample, that lower edge dips below the break-even line, which means a genuinely winning player can still finish the sample in the red purely through variance. That is not a flaw in their game; it is the expected behaviour of the maths.

Reading the two intervals together gives a practical sense of what is normal. Seven outcomes in ten fall inside the narrower 70 percent band, so results in that range should never feel surprising. Nineteen in twenty fall inside the 95 percent band, so an outcome near its edges is rare but entirely possible. Only results well outside the 95 percent band are a genuine signal that the win rate or standard deviation used was wrong. The wider story of how variance separates skill from luck over time, and how the bb/100 win rate is measured in the first place, sits behind every figure this tool produces.

04What a normal downswing can look like

The hardest part of variance is accepting how rough a perfectly normal stretch can feel. A genuine winner can go tens of thousands of hands below their expected line, and in a big or high-variance format can sit below break-even for far longer than instinct says is fair. None of that means the edge has disappeared - it means the sample is still small enough for noise to dominate the signal.

  • Winning players still lose for stretches. A positive long-run win rate says nothing about any given month.
  • Bigger swings, wider fields. The higher the standard deviation, the deeper and longer the realistic dips.
  • Short samples lie. A few thousand hands can look far worse - or far better - than the true long-run expectation.

A run only becomes a real warning sign when it falls well outside the 95 percent band above. Inside it, even the grim lower edge is simply the maths doing what it always does.

05What this calculator helps you judge

Read together, the expected line and the two intervals answer the questions that actually keep players up at night:

  • Is this downswing still normal, or a genuine sign that something has changed?
  • How wide can my results realistically be over a sample of this size?
  • Is my win rate strong enough that the lower edge of the band stays survivable?
  • How much of my current result is signal, and how much is still pure noise?
// faq
How does the poker variance calculator work? +

It uses two formulas. Expected profit in big blinds equals your win rate times the number of hands divided by 100. The standard deviation over the whole sample equals your per-100 standard deviation times the square root of hands divided by 100. From the expected value and that spread it derives a 70 percent and a 95 percent confidence interval, and charts the band around the expected line.

What standard deviation should I use for poker? +

For full-ring and six-max no-limit cash games a standard deviation of roughly 80 to 100 big blinds per 100 hands is typical, which is why the calculator defaults to that range. Looser, more aggressive games and formats like pot-limit Omaha run higher, often well past 100, while tight passive games can sit a little lower.

How big can a normal downswing be? +

Larger than most players expect. Because the spread grows with the square root of hands, even a solid winner will spend long stretches below their expected line and can run below break-even over tens of thousands of hands. The 95 percent interval shows the range that contains nineteen out of twenty realistic outcomes, and its lower bound is often deeply negative.

Does a high win rate remove variance? +

No. A higher win rate lifts the whole confidence band, so the expected outcome improves and the chance of finishing down falls, but the width of the band - the variance itself - is set by the standard deviation and the sample size, not by the win rate. Skill shifts the average; it does not flatten the swings.

Can a winning player have a long losing stretch? +

Absolutely, and it is more common than most players believe. Because the spread grows with the square root of the sample, a genuine winner can spend tens of thousands of hands below their expected line and even below break-even. As long as the result stays inside the 95 percent band the calculator draws, a losing stretch is normal variance rather than evidence of a leak.

How many hands make variance less dominant? +

There is no hard cutoff, but the edge only starts to clearly overwhelm the noise once the sample runs into the hundreds of thousands of hands for cash games. Below that, the confidence band is wide enough that two players with identical win rates can post very different results. More volume narrows the band relative to the expected line, which is why long-term samples are the only fair measure of a win rate.