How big are downswings, really?

Most players underestimate variance until they've lived through a big downswing. Let's put numbers on it.

Take a solid 6-max player: 5 bb/100 true win rate and a standard deviation of 100 bb/100, which is typical for no-limit hold'em cash games.

Over 100,000 hands

• Expected result: +5,000 bb
• 95% of the time, the result lands somewhere between about −1,200 bb and +11,200 bb
• Chance of being down after 100k hands: about 6%

So a genuinely winning player has about a 1-in-17 chance of showing a loss over a sample most people would call large.

Downswings inside the sample

The final number isn't the scary part. In simulations of this player over 100k hands:

• About 9 in 10 samples contain a downswing of at least 1,500 bb (15 buy-ins) from a previous high
• About 6 in 10 contain one of at least 2,000 bb (20 buy-ins)
• The typical worst downswing is around 2,200 bb

That's normal for a winner, not a sign that something is broken.

Run your own numbers in the variance calculator →

What this means for you

Your tracker's win rate is an estimate, not a fact. After 100k hands, a measured 5 bb/100 is consistent with a true win rate anywhere from about −1 to +11 bb/100. Keep that range in mind before you move up, move down, or overhaul your game.

Judge decisions, not results. Over short samples, results mostly measure variance. A database review looks at how you play each node, compared with a baseline, which doesn't have to wait a million hands to be meaningful.

Bankroll for the downswing you'll actually get. With the numbers above, a 30 buy-in bankroll (3,000 bb) gives a risk of ruin of about 5%, if you never move down. Most serious players want much less than that, which is why moving down during a downswing is part of good bankroll management.