What Are They Holding?
Given a number of players, what’s the chance someone at the table has at least a pair, two pair, trips, or better by the river? These two models tell very different stories.
Based on 100,000 simulated 6-player hands. About 23.34% of trials had only one player left (an uncontested pot).
Why there are two models, and which one to believe
Almost every “odds someone has a flush” figure you will find online quietly assumes nobody folds. It is the easier question: deal every seat two cards, run the board out, and count. That model is on this page as Everyone sees the river, and it is genuinely useful — because it is the one that reproduces the textbook five-card hand frequencies, which is how you know the simulation is not lying to you.
It is also not the game you play, and the difference runs the opposite way to most people’s intuition. At 6 players, someone holding at least two pair by the river is 83.9% if nobody ever folds, but only 59.0% once they do — folding makes the table less dangerous, not more.
That is worth sitting with, because the obvious argument points the other way: the hands that fold are the weak ones, so the hands that survive to a showdown are individually much stronger than a random two cards. Both halves of that are true. But the question here is whether anybody has a big hand, and folding removes most of the anybodies. Fewer opponents beats stronger opponents, and by a wide margin — which is why a pot that folds around to three players is a different proposition from one where all 6 call.
Both numbers are correct answers to different questions. The one you want at the table is the second.
What the table size actually does
More players means more chances that somebody has something, and the effect is larger than most people carry around in their heads. Under the realistic model, the chance that at least one opponent holds two pair or better is 47.4% heads-up against 64.0% nine-handed. A top pair that is comfortably ahead heads-up is a hand you have to think much harder about at a full table — nothing about your cards changed, only the number of people who had a chance to beat them.
The same pressure shows up in how often a pot is contested at all. Step the table size up and the share of hands that end with a single player left falls steadily — more seats means more chances that at least one of them wakes up with a hand worth playing.
The folding model has one more consequence worth naming: at 6 players, 23.34% of hands end with only one player left, uncontested. Those pots never reach a showdown at all, which is why the realistic figures here describe the hands that do.
Reading the chart properly
The bars are cumulative — “at least” rather than “exactly”. The bar for Flush at 6 players (12.3%) means somebody has a flush or better, so it includes the full houses and quads above it. Reading it as “exactly a flush” overstates how often you are up against precisely that hand and understates how often you are drawing dead to something bigger.
One thing this page deliberately does not tell you is what to do about it. Knowing an opponent probably holds three of a kind or better is only half a decision — the other half is what you hold and what it costs to find out. That is the outs calculator and the preflop equity tables.
Both models are Monte Carlo rather than exact — there is no closed form for “somebody at the table has at least X” once folding is involved. Each figure here comes from 100,000 simulated hands, which is enough to settle the first decimal place and not much beyond it. The methodology page sets out what the folding model assumes, and it assumes plenty.