Beyond RangeBeyond Range · The Verdict
Are you a winning player, or just running good?
Your graph is not your report card.
You came here for a verdict on whether you’re any good. Here’s the honest one: your results aren’t allowed to say yet. Watch why.
Most players are smaller winners than they think. The verdict on whether you’re a winner never moves — that’s read from your results alone. But a smaller assumed edge makes your swings look more normal, and the road to certainty far longer.
The tax that quietly lengthens every road to certainty.
Every number here is closed-form or seeded Monte Carlo, unit-tested in your browser — no fitted curves, nothing stored. The math.
Buy-ins won / lost · over the same hands, at the same skill · tap to skip
Where your true winrate could really be · narrowing as you play
Go deeper
Two players sit down for a month. One truly wins 5bb/100; one truly loses 2bb/100. Run them side by side and the loser will out-earn the winner more often than you’d ever guess — for tens of thousands of hands at a stretch. A good month can lie exactly as loudly as a bad one.
Real graphs with a lot of stacks going in run a little wilder in the short term — so a swing like this is, if anything, slightly more common than shown here, never less. The model never flatters your luck.
The longer read: why results lie, and what to measure instead.
You’re responsible for what you could reasonably know — not for what the cards were.
The book of this tool — The Downswing: Variance, and the Verdict You Can’t Yet Reach. The subtitle is this tool’s premise, written whole: what to do on the nights the number can’t judge you yet. Read it online or take the PDF.
Related reading: how long downswings actually last · running bad, or playing bad?
On a phone: press and hold the image above to save it.
The occasional essay on poker and truth. No spam, no funnel.
Are you actually winning at poker — or just running good?
Every serious player asks it eventually, usually in the middle of a downswing: am I a losing player, or am I just running bad? It feels like a question about your soul. It’s really a question about sample size — and it has a real answer, usually an uncomfortable one. The tool above computes that answer honestly. Here is what it’s doing, and why the honest answer is so often “you can’t know yet.”
Am I a losing player, or just running bad?
Your results are two things stacked on top of each other: your true edge, and noise. Your edge is what you’d earn per hand if you could play forever — your real winrate, buried somewhere inside you that you never get to see directly. Noise is everything else: the coolers, the two-outers, the stretch where every flip lost. What lands on your graph is edge plus noise, and over any normal sample the noise is far larger than the edge. So your graph is mostly a picture of your luck, wearing a thin coat of your skill.
That is why “am I a loser, or unlucky?” is so hard to feel from the inside — the two look identical across tens of thousands of hands. The only way to separate them is a colder question: given how much I’ve played, could a genuine winner have results this bad by luck alone? Almost always the answer is yes, and The Verdict puts a number on exactly how often.
Is a 15 buy-in downswing normal?
Yes — not just possible, ordinary. A solid 5bb/100 winner in a typical 6-max game hits a downswing of 15 buy-ins or worse somewhere inside 20,000 hands about one time in five, and plays no differently doing it. Stretch the sample and the number that would shock you keeps happening: downswings of 20, 30, even 40 buy-ins are not evidence of a broken game or a broken player. They are the texture of a small edge sitting underneath a large standard deviation. The size of a downswing tells you almost nothing about whether you’re good. It mostly tells you how many hands you’ve played.
Why your results lie to you
Here is the mechanism in one sentence: your edge grows in a straight line with the hands you play, but your variance only grows with the square root. Play a thousand hands at 5bb/100 and you’ve earned fifty big blinds on average — but the swing around that number is many times larger. The edge is real the whole time; it’s just a whisper, and variance is shouting over it. Only as the sample stretches into the tens and hundreds of thousands does the whisper start to win, because the straight line keeps climbing while the square root crawls. Every losing month a real winner endures is that arithmetic, seen from the inside.
How many hands until you can trust your winrate?
Far more than anyone wants to hear. To be 95% sure a 5bb/100 winner is actually beating the game — not merely “probably not losing” — takes on the order of 90,000 hands in 6-max. And the part that stings: the number scales with the square of your edge. Halve your true winrate and you don’t need twice as many hands to prove it, you need four times as many. A genuine 2bb/100 winner — a common, real grind — needs over half a million hands before their graph is allowed to call them a winner. Most players draw a conclusion after a few thousand. The signal arrives tens or hundreds of thousands of hands later. That gap, between when you decide and when you could actually know, is where most tilt lives.
Running bad or playing bad — how to actually tell
If your results can’t tell you, what can? Your decisions. Variance owns your outcomes over any short stretch, but it never touches the quality of the choice you made with the information you had. A good fold that got coolered was still a good fold; a bad call that spiked the river was still a bad call. The only reliable signal in the short run is whether you can defend your decisions against what you knew at the time — not against what the deck did. That is the whole reason results-oriented thinking is a leak: it scores the decision by the outcome, and the outcome is mostly noise. If you want to improve this month, stop auditing your graph and start auditing your reads.
What “not enough hands to know” actually means
For most players, most of the time, the honest verdict is “you can’t tell yet” — and that is not a dodge, it’s the finding. Here is the standard behind it. Take your observed winrate and put a 95% confidence interval around it: the range your true winrate is very likely to sit inside, given how much you’ve played. If that whole range is above zero, the data has earned the word “winner.” If it’s entirely below zero, “loser” is real and worth fixing. But if the range straddles zero — and after a few tens of thousands of hands it almost always does — then the same sample is consistent with both a solid winner and a small loser, and no amount of staring at your graph resolves it. That straddle of zero is the whole answer. The Verdict shows you the exact interval, so the ruling is something you can see, not something you’re asked to trust. Crucially, that ruling is read from your results alone; changing the “assumed edge” control never moves it, because you can’t argue yourself into being a winner — only play yourself there.
What The Verdict assumes — and what it deliberately won’t do
Honesty about the model matters. The Verdict treats a cash-game session as a normal distribution around your true winrate, using a standard deviation typical for your game — higher for heads-up and online, lower for full-ring and live. Real short-run poker has slightly fatter tails than a clean bell curve because of stack-off variance, which means extreme swings happen a little more often than the tool shows, never less. It errs toward “your run is normal,” never away from it.
It does not model tournaments. MTT results follow a different, lottery-shaped distribution, and a normal-curve verdict on them would simply be wrong — so the tool refuses to give one. It is a cash-game instrument. It also won’t size your bankroll or compute risk of ruin; if you want those raw numbers, the Variance & Risk of Ruin calculator does exactly that. The Verdict answers a narrower, more personal question: can your results judge you yet?
Common questions
Is my poker downswing normal? Almost certainly. A genuine 5bb/100 winner hits a 15-buy-in downswing somewhere inside 20,000 hands about one time in five, and plays no differently doing it. The size of a downswing is mostly variance, not a leak.
How do I know if I’m a winning poker player? By the only honest standard: when the 95% confidence interval on your observed winrate no longer includes zero. Until that interval clears break-even, your results genuinely cannot say — no matter how far up or down you are.
How many hands until I know my true winrate? Far more than most think. Confirming a 5bb/100 edge takes roughly 90,000 hands in 6-max; a 2bb/100 edge takes over 500,000. Certainty scales with the square of your edge.
Running bad or playing bad — which is it? Your results can’t tell you over any normal sample; only your decisions can. Judge the choices you made against what you knew at the time, not against what the cards did.
Related reading: Poker Variance and Downswings · The Cult of Results Is Lying to You · Variance & Risk of Ruin calculator