What Is a Good Poker Winrate — and What Does Yours Really Mean?
bb/100 describes results. It does not name the decision that caused them, and it is a poor leak detector on its own.
Winrate tells you what happened. It does not identify the decision that caused it.
REQUEST EARLY ACCESS · ALPHA IN PREPARATION
REQUEST EARLY ACCESS · ALPHA IN PREPARATION
RESULT METRIC · NOT A LEAK DETECTOR
How bb/100 is calculated
Observed bb/100 rescales net big blinds so results can be compared across different hand counts.
(net big blinds won / hands played) × 100
The formula is a result identity. It does not name a decision, a leak, or a true skill rate.
Why observed winrate is noisy
All-in runouts, cooler stacks, and short samples move the number even when the underlying decisions have not changed.
- Observed results include realized luck in all-in pots.
- A few large pots can dominate a small file.
- The same decisions can print different bb/100 over different windows.
What all-in adjusted bb/100 means
All-in adjusted changes the treatment of specific all-in outcomes. It is another estimate, not a complete map of every decision.
Observed
Actual table result.
All-in adjusted
Changes treatment of specific all-in outcomes.
It can:
- Reweights some all-in showdowns
- Reduces one source of realized luck
It does not adjust:
- Every bluff
- Every fold
- Every thin value bet
- Every missed opportunity
- All strategic variance
Why EV-adjusted isn't “true winrate”
All-in adjusted is not true winrate. It does not repair every form of uncertainty, and it is not a complete measure of decision quality.
How stakes and rake affect comparisons
A “good” winrate is not portable. Stake, game environment, rake, player pool, sample, table format, and selection all change what a number can mean.
- Stake
- Game environment
- Rake
- Player pool
- Sample
- Table format
- Selection
This page does not publish a universal crusher table. Those thresholds hide the conditions that make the number comparable.
When negative winrate deserves investigation
Negative bb/100 is a result, not a player verdict. Use it as a reason to inspect sample, position, aggregates, and repeating decisions.
- Negative result
- Check sample / volatility
- Check positional results
- Check aggregate signals
- Check repeating decision patterns
Analyze is the current entry point for that investigation. Full automated diagnosis is still being built.
Why winrate cannot identify a leak
A result number collapses many decisions into one outcome. It cannot tell you which pattern to study next.
Decision-level re-measurement is the intended verification target. Short-term winrate does not verify improvement, and fixing one pattern is not claimed as a stated bb/100 lift.
Winrate, asked directly
Is all-in adjusted bb/100 my true winrate?
No. It is another estimate with its own blind spots. It is not a complete measure of decision quality.
Can you improve my winrate by a stated number?
We will not claim a causal short-term winrate lift. The product measures whether targeted decisions improve on new hands.
Look behind
the result number.
Winrate is the scoreboard. The next useful object is a defined decision with evidence.
REQUEST EARLY ACCESS · ALPHA IN PREPARATION
No claimed causal winrate lift. Decision-level re-measurement is the intended check.