Your sample
Big blinds won per 100 hands, from your tracker.
Your total sample. Live players: ~25–30 hands an hour.
How wild your swings are. Don’t know it? ~100 is typical for online 6-max No-Limit Hold’em.
What the data can support
95% confidence range for your true win rate
−3.8 to +13.8
bb/100 — where your real rate most likely sits.
- Observed win rate+5.0 bb/100
- 70% range (more likely)+0.4 to +9.6
- Margin of error (95%)± 8.8 bb/100
- Chance you’re a long-term winner87%
The 95% range still straddles zero, so this sample cannot yet prove you are a winning player — your true rate could plausibly be a loss. This is exactly the gap most trackers paper over.
How this is calculated
Your win rate is an average, and every average from a sample has a margin of error. The standard error of a win rate measured over n hundred-hand blocks is SE = SD / √(hands / 100). The 95% confidence interval is your observed rate ± 1.96 × SE; the 70% interval uses ± 1.036 × SE.
The “chance you’re a long-term winner” is the normal-distribution probability that your true rate is above zero, given this sample (Φ(win rate / SE)). It’s an approximation that assumes roughly normal results and a known standard deviation, so treat it as a guide, not a verdict. The headline is simple: smaller samples mean wider ranges, and wide ranges can’t prove much.
Why a single win-rate number lies to you
A win rate like 5 bb/100 sounds precise, but it’s a measurement, and measurements have error bars. Over a small sample those error bars are enormous — your 5 bb/100 could comfortably be the face of a true rate anywhere from a healthy win to an outright loss. That’s not a flaw in your play; it’s variance, the random noise baked into every poker result.
The calculator turns your sample into an honest range instead of a false point. The wider the range, the less you actually know. As your hands climb, the range tightens — which is the entire reason StackWise withholds a headline win rate until your data can stand behind it.
If the 95% range still crosses zero, you cannot yet claim to be a winning player — no matter how green the number looks. Read how many hands it really takes to close that gap.
Stop guessing. Let StackWise track the range for you.
StackWise pairs every win rate with its confidence range and a sample-size label, automatically, for every session you log. Free to start, no account.
Sources & method
Confidence intervals use the standard error of a mean, SE = SD / √(hands ÷ 100), with z = 1.96 (95%) and 1.036 (70%). Typical standard-deviation values and sample-size guidance from ThePokerBank, Primedope, and Poker Trainer.