Variance is the reason you can do everything right and still lose for weeks. It's the invisible force that separates your short-term results from your true EV. Understanding it isn't just mathematically useful — it's essential for keeping your mental stability and bankroll intact when the cards aren't falling your way.
In this guide, you'll see what variance is, how to measure it, what a "normal" downswing looks like, and how to manage it without destroying your bankroll or your mindset.
1. What is Variance?
Mathematically, variance measures how much actual results deviate from expected value (EV). In poker, it's the difference between what you should win and what you actually win in the short term.
The key tool is standard deviation (SD): how many bb/100 hands your results can vary from EV.
Typical 6-max cash game player
Winrate
+3 bb/100
SD
~100 bb/100
This means that in a 100-hand session, the "normal" result (68% of the time) is in the range of −97 to +103 bb. You can lose 97bb while playing perfectly.
At 95% probability (2 standard deviations), the range is −197 to +203 bb in a single 100-hand session. This isn't extraordinary bad luck — it's normal mathematics.
Why it happens
Poker combines EV decisions with random outcomes on every hand. Even if you make the correct decision 100% of the time, the individual result of each hand is uncertain. Over time, results converge to true EV. In the short term, variance dominates.
2. How Many Hands Do You Need to Know If You're Good?
Variance decreases with volume. The more hands you play, the closer your real results get to your true EV. The formula is:
95% CI = Winrate ± 2 × SD / √(N/100)
(N = number of hands)
For a player with 3 bb/100 winrate and 100 bb/100 SD:
| Hands played | Possible real range (95%) | Valid conclusion? |
|---|---|---|
| 1,000 | −5.3 to +11.3 bb/100 | No |
| 10,000 | +0.7 to +5.3 bb/100 | Partially |
| 50,000 | +2.1 to +3.9 bb/100 | Fairly reliable |
| 100,000 | +2.4 to +3.6 bb/100 | Very reliable |
The uncomfortable conclusion: with 10,000 hands you still can't be statistically certain of your winrate. A −1 bb/100 player can show positive results for thousands of hands from pure variance.
This also works in reverse: a +5 bb/100 player can show negative results for 20,000 hands. Variance doesn't distinguish between good and bad players in the short run.
3. Downswings, Coolers and Bad Beats
How large can a normal downswing be?
A downswing is a streak of results below EV. With an SD of 100 bb/100, the expected maximum downswing over N hands can be estimated as:
Max downswing ≈ 2 × SD × √(N/100)
Over 10,000 hands: 2 × 100 × √100 = 2,000 bb = 20 buy-ins
A 20 buy-in downswing is statistically expected. It's not a sign you're playing poorly.
Cooler vs Bad Beat: the distinction that matters
Not every loss is the same. Identifying the type changes how you react:
Cooler
Both players made correct decisions. Example: KK vs AA preflop. You lost without making any mistake. Pure variance.
Bad Beat
You were the clear favorite (80%+) and lost. Your decision was correct and EV was on your side. The adverse result was variance. Example: you lost a hand with 90% equity on the turn.
Mistake disguised as bad beat
You think it was variance but it was actually an incorrect decision — calling with insufficient equity or bluffing without fold equity. The adverse result was the mathematically expected consequence, not bad luck.
Only genuine bad beats are variance. Mistakes are negative EV that eventually gets collected.
4. How to Manage Variance
Bankroll: the shield against variance
Bankroll management exists precisely to survive statistically normal downswings without going broke.
| Format | Minimum buy-ins | Why? |
|---|---|---|
| 6-max cash game | 20–30 BI | Moderate variance |
| PLO (Pot Limit Omaha) | 40–50 BI | Very high variance |
| MTT (tournaments) | 50–100 BI | Downswings of hundreds of tournaments are normal |
| Sit & Go | 30–50 BI | Intermediate variance |
Volume: the only real solution
The law of large numbers guarantees that with enough hands, results converge to EV. You can't speed up this process, but you can make sure you're there to benefit from it by playing sufficient volume with quality decisions.
Mental game during a downswing
- Review handsAnalyze objectively. Are you losing because you're playing badly or because variance is adverse? They're different problems with different solutions.
- Watch the EV lineTrackers like PT4 or HM3 show your EV line (all-in EV) separate from actual results. If EV is climbing but results are falling, you're playing well and variance is temporary.
- Accept the inevitableA 15 buy-in downswing in a month is not a sign you're playing badly. It's statistics. The right question isn't "why is this happening to me?" but "am I making +EV decisions?"
Conclusion
Variance isn't your enemy. It's simply the mathematical nature of the game. Players who understand variance don't tilt when they lose, don't move down in stakes in a panic after a normal downswing, and don't confuse results with quality of play.
Play well, maintain your bankroll, accumulate volume, and let the math do its work.