Poker variance explained: how to actually measure it

Poker variance is the gap between how well you play and how your results actually look on any given week. Skill decides where you land in the long run. Variance decides how wild the ride is on the way there - and unlike luck, it is something you can measure.

Most players treat variance as a vibe - good run, bad run, "the deck hates me." But it is a real statistical quantity with real numbers attached. Once you can put a figure on your poker variance, downswings stop feeling like a verdict on your game and start looking like exactly what the math predicted. Here is how it is actually measured, from the basic formula to the Monte Carlo simulations that power a modern variance calculator.

What poker variance actually is

In statistics, variance measures how far results spread out around their average. In poker, that average is your expected value - your true win rate - and variance is how far individual sessions scatter above and below it. A low-variance game keeps results tight to your EV. A high-variance game throws huge wins and brutal losses around the same long-term number.

Two players can share an identical win rate and have completely different experiences getting there. That spread is variance, and it is the reason a single losing month tells you almost nothing about whether you are good. For the emotional side of riding it out, our guide to handling downswings goes deeper - this article is about the numbers underneath it.

Standard deviation: the number behind the swings

The everyday measure of poker variance is standard deviation - the square root of variance, expressed in the units you actually care about. For cash games that unit is BB/100 (big blinds per 100 hands). A typical no-limit hold'em winrate comes with a standard deviation somewhere around 80-100 BB/100, which is why a 5 BB/100 winner can still be stuck after tens of thousands of hands.

Tournament players measure it in buy-ins instead, paired with ROI. Because MTTs are so top-heavy - most of the prize money sits at the very top - their standard deviation dwarfs cash games. The same edge produces far wider swings, which is exactly why tournament bankroll requirements are so much steeper.

Why sample size changes everything

Here is the part that breaks most people's intuition: variance does not shrink as you play more, but its impact relative to your edge does. Over 1,000 hands, variance overwhelms skill completely. Over a million, skill almost always wins out. That is why a weekend, a session, or even a month proves nothing - and why you need a serious sample before your expected value shows up in your actual bankroll.

The math is simple in spirit: your edge grows linearly with volume, but the swings only grow with the square root of it. Play four times as many hands and your expected profit quadruples while your typical swing only doubles. Time is on the skilled player's side - but only if the bankroll survives long enough to get there.

Monte Carlo simulation: the modern way to model variance

You can calculate variance with formulas, but the clearest way to see it is a Monte Carlo simulation. Named after the casino, it is a method borrowed from finance and physics: instead of solving an equation once, you let the computer play out your inputs at random thousands of times and watch the distribution of results that emerges.

Feed in your win rate, your volume and the shape of your games, run it 10,000 times, and you get something no formula gives you - a picture of every plausible version of your year at once. The lucky paths, the disastrous ones, and the dense cluster in the middle where most outcomes actually land. That is precisely what Pokercode's free variance calculator does, and we break down how to read its charts in our companion piece on the poker variance calculator.

Why simulation beats a single formula

A formula can hand you one standard deviation figure. A simulation shows you the whole tail - how often the worst case actually happens, how deep a downswing can plausibly run, and what fraction of outcomes would have busted a given bankroll. For decisions like "is my roll big enough for these stakes," the distribution matters far more than the average.

Risk of ruin: where variance meets your bankroll

Risk of ruin is variance made personal. It answers one question: given your edge, your swings and your bankroll, what are the odds you go broke before the long run rescues you? A genuine winner with a bankroll that is too small can still have an uncomfortable risk of ruin - the edge is real, but the variance gets to the money first. Lowering it is simple in principle: win more, or carry more buy-ins. A good bankroll management plan is really just risk of ruin kept to a number you can sleep on.

Other ways players estimate variance

Monte Carlo is the cleanest, but it is not the only approach:

• Tracking software - tools like Hold'em Manager and PokerTracker calculate your standard deviation and confidence intervals straight from your hand history.
• Confidence intervals - a way of saying "my true win rate is very likely between X and Y," which gets tighter the more you play.
• Downswing simulators - purpose-built to answer "how long and how deep can a downswing get at my win rate."
• Manual records - for live players with no tracker, a disciplined results log is the only honest source of a sample.

Turn the math into fewer leaks

Understanding poker variance is the part that keeps you sane. Shrinking the variance you can actually control - the repeated preflop, ICM and final-table decisions that set your real win rate - is the part that makes you money. Model the swings, then go fix the inputs.