Pokeriad / Learn / Luck or skill
Updated November 2026
Poker is both: one hand or one night is mostly luck, and over thousands of hands skill decides who ends up ahead. In the best-known study, players picked out as skilled before the 2010 World Series of Poker earned 30.5% on their entry fees, while everyone else lost 15.6%.
| Study | Data | Finding |
|---|---|---|
| Levitt and Miles, 2011 | 2010 World Series, 32,496 players | Skilled group +30.5%, the rest −15.6% |
| Potter van Loon and others, 2015 | 76.9 million online hands | Skill beats luck after about 1,500 hands |
| Fiedler, 2014 | 51,761 online players | A typical losing player: after 767 hands |
| Meyer and others, 2013 | 300 players, 60 hands | Over 60 hands, the cards decided it |
| Brown and Sandholm, 2019 | 10,000 hands against professionals | A computer program beat them |
Economists Steven Levitt and Thomas Miles used information from before the 2010 World Series of Poker to pick out 720 players as highly skilled, from published rankings, past titles and the previous year's big earners. Then they checked how those players did across all 57 tournaments.
The 720 averaged a return of 30.5% on their entry fees, more than $1,200 per player per event. Everyone else averaged −15.6%, a loss of more than $400. If poker were pure luck, a list drawn up in advance couldn't have done better than the rest of the field.
It depends how far apart the players are. Rogier Potter van Loon and colleagues studied 76.9 million online hands and found skill outweighs luck after about 1,500 hands, using the rule that the better player has to be ahead 3 times in 4.
That's 19 to 25 hours at one table, and it's the figure for comparing the best 1% of players with the worst 1%. Compare the top and bottom 10% and it takes 6,500 to 7,300 hands. Ingo Fiedler of the University of Hamburg put it at 767 hands for the typical player, who is a losing one.
Yes, which is hard to explain with luck. In that same online data, players were ranked on their first six months and then followed through the next six. Those in the top 10% stayed there 20.7% of the time, where chance alone would give 10%.
The effect is real and also modest. The researchers could explain only a small part of the differences between individual players from past results, because luck covers most of it for anyone who hasn't played many hands.
Yes, and it deserves a fair hearing. Gerhard Meyer and colleagues sat 300 people down for 60 hands of computer poker and rigged the deal, so some got better cards than others. The skilled players didn't finish with more money than the average ones.
The cards decided it, and the authors concluded that poker should be regarded as a game of chance, at least under those conditions.
Sixty hands is the catch. The other studies show that luck is expected to dominate that short a game. The skilled group did do one thing better: they lost less when they were given bad cards.
It wins, and the cards can't be the reason. In 2019 a program called Pluribus played 10,000 hands of six-player no-limit hold'em against professionals, each of whom had won more than $1 million at poker. It won 4.8 big blinds per 100 hands.
An earlier program, Libratus, beat four top specialists at two-player poker over 120,000 hands. Those programs get the same cards as anyone else, so their results have to come from the decisions.
Big blinds per 100 hands, with the standard error.
The cards are random every time, and your decisions aren't. A single hand can reward a bad decision, as this $1/$2 hand shows, but the sum behind that decision comes out the same way every time you make it.
You have a flush draw and your opponent bets $40. You need to win 33% of the time to call, and you'll hit 19.6% of the time. Your call loses $16.52 on average, and you make it anyway.
A heart comes and you win $80. It feels like a good decision.
The luck evens out and the decision is what's left. The same thing happens in the other direction to a good decision that loses.
A player who wins 5 big blinds per 100 hands with a swing of 90 is ahead after 100 hands only 52 times in 100, barely better than a coin. After 100,000 hands it's 96. Skill is present in every hand, and it takes that long to rise above the noise.
Poker variance shows what those swings look like.
Without anyone showing their cards. A 2009 analysis of 103 million online hands by the software firm Cigital found 75.7% ended with everyone folding to one player, so the best cards didn't decide them.
Treat that number with some care. The work was done under contract to the company that runs the site those hands came from, and the report itself says it doesn't measure the effect of luck.
A federal judge in New York, Jack Weinstein, ruled in 2012 that skill predominates in poker and overturned a conviction for running a game.
The Second Circuit reversed that in 2013, holding that the question of skill or chance didn't apply under the federal law involved. The appeals court never decided whether poker is a game of skill. Laws differ from one place to another, so check yours before you assume anything.
Don't judge yourself by a night, a week or a month, because over those lengths luck can be louder than anything you do. Judge the decisions. If you make better ones than the people you're playing, the results will show up eventually, though you may need more hands than you expect to see them.
It doesn't mean you'll be one of the winners. Can you make money playing poker? looks at how few people are.
Twelve spots in three rounds.
In one hand, luck. In tens of thousands of hands, skill. Somewhere between those is the point where they're about equal, and for players far apart in skill it comes early.
Yes, in the everyday sense and in the law of most places. You're risking money on an uncertain result, however good you are.
In one hand or one night, yes, and it happens all the time. Over a long match, the professional wins nearly every time.
Often because they play too big for their bankroll, so normal swings wipe them out before skill has time to show. Bankroll management covers that.
Figures for each study come from the papers linked next to them. The simulator treats each player's results as normal, with a swing of 90 big blinds per 100 hands, and works out how often the better player is ahead of the other.
The 75% line is the rule used in the online study: skill counts as outweighing luck when the better player is ahead 3 times in 4. The 20.7% figure ranks players by that study's measure adjusted for how much their results swing.
Send them the studies.