The Truth About Poker Variance

In statistics, variance measures how spread out results are around the expected value. Standard deviation — the square root of variance — is the more intuitive number because it lives in the same units as your results. In poker, we typically express standard deviation in big blinds per 100 hands (bb/100).

For No-Limit Hold'em cash games, a typical standard deviation is 80–120 bb/100. A very tight, conservative player might run closer to 70 bb/100, while an aggressive LAG style or a high-SPR deep stack game can push standard deviation well above 100 bb/100. For purposes of this article, we'll use 100 bb/100 as our baseline — a reasonable approximation for most 6-max or full-ring cash games.

Now apply that to a concrete player. Imagine a

Scale that up to 1,000 hands. Expected profit: 50 bb (

The first question every player in a downswing asks is: when does this end? The honest mathematical answer is: it ends when the sample size is large enough for your edge to assert itself — and that takes longer than almost anyone expects.

For our 5 bb/100 player with 100 bb/100 standard deviation, the “breakeven point” — the number of hands needed before there's a 95% chance of being in profit — is approximately 10,000–12,000 hands. At a typical live cash game pace of 25–30 hands per hour, that's 350–480 hours at the table. For a player logging two sessions per week of 3 hours each, that's over a year of play before variance is “resolved” with high confidence.

Online players move faster — 60–100 hands per hour on a single table, more with multi-tabling. A dedicated online grinder might log 1,000 hands in a long session. That compresses the timeline significantly: 100 sessions of 1,000 hands each gets them to the 95% confidence threshold. Still months of play, but vastly faster than live.

The important insight is that downswings are not exceptional events for winning players — they are mathematical certainties. A player winning at 5 bb/100 will experience at least one 5-buy-in downswing with near 100% probability over a standard poker career. Multi-buy-in downswings lasting thousands of hands are not evidence that you have lost your edge. They are evidence that you are playing poker.

The most dangerous thing about variance is that it provides cover for leaks. “I'm just running bad” is both the most common true statement and the most common excuse in poker. Knowing which one applies to you requires honest analysis, not intuition.

The key diagnostic is your all-in equity adjusted win rate, sometimes called your “EV adjusted” or “luck-adjusted” line. Tracking software like PokerTracker and Hold'em Manager calculates this by evaluating your equity in every all-in situation at the moment chips go in, rather than crediting the actual outcome. If your equity-adjusted line is trending upward but your actual results are down, you are running below expectation — genuine variance. If both lines are flat or declining, you have leaks.

Beyond the EV line, there are specific hand categories worth auditing. Are you losing money on hands where you should be printing? Check your win rate with strong made hands at showdown, your c-bet success rate, and your big blind defense outcomes. If your strong hands are losing to unlikely runouts at a rate that's clearly above expectation, variance is the likely culprit. If your continuation bets are getting called and outdrawn at normal frequencies but you're still losing, your sizing or range construction may be the problem.

One practical heuristic: if your downswing is under 5 buy-ins at your typical stakes, it is almost certainly pure variance. Between 5 and 10 buy-ins, conduct a leak audit but do not panic. Above 10 buy-ins, you should rigorously review your database — not because the downswing proves you are losing, but because a sample of that size contains enough information to identify whether leaks are contributing. The downswing might still be variance, but it is now large enough to take seriously as a diagnostic opportunity.

The purpose of a bankroll is not to fund playing — it is to absorb variance long enough for your edge to manifest. A bankroll that is too small will bust before the law of large numbers can work in your favor, even if you are a genuinely winning player. Sizing your roll correctly is therefore not conservative — it is strategic.

The conventional recommendation of 20 buy-ins for cash games is a floor, not a ceiling. That figure assumes a reasonable win rate, controlled variance, and a willingness to move down in stakes if the roll is threatened. For players with lower win rates, higher variance play styles, or a reluctance to drop down, 30–40 buy-ins is more appropriate. Tournament players face even steeper requirements: 100+ buy-ins is not unreasonable for players whose primary game is MTTs.

Understanding variance intellectually is necessary but not sufficient. The real challenge is applying that understanding when you are three buy-ins deep in a session, your last four sets have been cracked, and the recreational player across the table has been sucking out on you all night. In those moments, the math feels abstract and the pain is very concrete.

The most effective mental game technique for variance is process orientation. Instead of measuring your session by dollars won or lost, measure it by decision quality. Did you find the right folds? Did you value bet thinly when the situation called for it? Did you stay composed when a cooler went against you? A session where you played perfectly and lost two buy-ins is a better session than one where you played sloppily and profited. The former contributes to your long-run edge; the latter obscures the leaks that will cost you later.

A second powerful tool is variance journaling. At the end of each session, note the hands where results diverged from your equity — the all-in flip you lost, the river suckout you took, the cooler where both players flopped top two. Seeing these events written down in a list makes them feel less personal and more statistical. It also gives you data for the next time you start to wonder if something unusual is happening: you can look at your journal and realize that variance events are, in fact, occurring at roughly normal frequency.

Finally, set explicit stop-loss limits for sessions and weeks, and honor them without exception. A common guideline is to stop a session after losing two buy-ins and to reassess if you drop more than five buy-ins in a week. These limits are not about giving up — they are about protecting yourself from tilt-induced compounding losses. Variance is painful enough on its own. Variance plus tilt is how bankrolls disappear. The best players in the world use stop-losses because they understand that no amount of mental strength fully immunizes you against the distorted thinking that comes with running badly for long stretches.