How Bad Can a Downswing Get? The Variance Math Every Serious Player Should Know

The downswing that makes you want to quit is almost never a leak. It is math. A break-even player, someone with a true win rate of exactly zero, will typically endure a 35 buy-in downswing across 100,000 hands, and a rough stretch takes them close to 70 buy-ins deep. That is not tilt or a broken game. It is what variance does to anyone who sits down. Understanding the size of normal swings is the difference between riding one out and blowing up your bankroll or your confidence at the worst possible moment. Pull your real win rate and standard deviation from Tracker+, then run the numbers yourself in our Variance Calculator as you read.

How deep does a normal downswing go?

Two inputs drive everything: your win rate in big blinds per 100 hands, and your standard deviation, which is how much your results bounce around that average. For six-max no-limit hold'em, standard deviation sits near 100 bb/100. Here is the worst downswing a player at each win rate should simply expect across 100,000 hands, from a Monte Carlo simulation of thousands of runs.

Read that again. A genuine crusher who beats the game for 8 bb/100 still runs 19 buy-ins under water in a normal stretch, and a bad month can push it past 30. The stronger your win rate, the shallower and shorter your swings, which is the real, unglamorous reason to keep studying: a higher win rate is the only thing that permanently tames variance. Tournament players have it far worse, since MTT standard deviation is several times higher and 100 buy-in downswings are routine. Set your format in the Variance Calculator to see your own numbers.

Why you cannot tell whether you are actually winning

Here is the part that quietly wrecks confidence. Take a true 5 bb/100 winner and ask what their results could look like after a given sample. The 95% range of observed win rates is enormous until you have hundreds of thousands of hands in.

At 100,000 hands, a real winner's results can still show anywhere from slightly losing to strongly winning. You literally cannot distinguish a small winner from a break-even player at that sample. And a genuine 5 bb/100 winner has a 13% chance of being in the red after 50,000 hands, and still a 5.7% chance after 100,000. Being stuck below zero across a big sample is not proof of a leak. It is a normal, expected outcome that happens to plenty of winning players. Only around half a million hands does the noise finally shrink enough to trust the signal.

Turn the math into a bankroll rule

Since swings are this deep, your bankroll has to survive them. Risk of ruin is the chance you ever go broke if you keep playing at the same win rate and standard deviation. For that same 5 bb/100 winner:

• 15 buy-ins: 22% chance of ruin. Far too thin.
• 20 buy-ins: 13.5%. Still uncomfortable.
• 30 buy-ins: 5%. A reasonable floor for a solid winner.
• 50 buy-ins: under 1%. Genuinely safe.

This is why the old 20 buy-in rule is optimistic for anyone but a crusher, and why 30 to 50 is the honest range. Note how sensitive it is to win rate: a break-even or marginal player has a risk of ruin approaching certainty at any bankroll, because there is no edge for the math to lean on. Size your roll for your real numbers with the Bankroll Calculator, and read more on bankroll management for serious players.

What to actually do with this

Three things. First, do not move down mid-downswing on emotion. A 25 buy-in slide is a normal event for a winner, not a signal that the games got tougher. If your expected value per hand is still positive, the swing is noise. Second, know your real standard deviation. Guessing 100 bb/100 is fine to start, but your true number comes from your database. Tracker+ records your results so your variance math reflects your actual game, not an assumption. Third, keep raising your win rate. The tables above make it concrete: every extra bb/100 you earn through study makes your downswings shallower, your bankroll safer, and your results readable sooner. Drilling real spots in GTO Ranges+ is variance reduction disguised as strategy work.

The takeaway

Poker variance is far more brutal than intuition suggests. A break-even player loses 35 buy-ins in a normal 100k-hand stretch, a solid winner can be down after 100,000 hands, and even crushers ride out 20 buy-in swings. None of that means you are playing badly. Learn the size of normal swings, keep a bankroll that can absorb them, and pull your own win rate and standard deviation into the Variance Calculator so the next downswing feels like weather you expected rather than proof that something is wrong.