If your aces just got cracked by kings, or your kings ran into aces, here is how unlucky you actually were: not very. Aces win 81.1% to 82.4% of the runouts against kings when the money goes in before the flop, and once you split the small number of chopped pots, their share is 81.3% to 82.6% depending on the suits. The kings were getting there a bit less than one time in five. One time in five is not a rare event.
The other version of this question, the one that starts “what are the odds” about somebody turning one chip into a mountain, usually has no answer at all. Odds need a fixed matchup: two hands, the cards to come, and nothing else moving. A run of hands over a night has none of that, so the number does not exist and nobody can compute it for you. What can be computed is the collision. At a nine-handed table you are dealt kings about once in 221 hands, and when you are, somebody else is holding aces about once in 26 of those times. That is the part of your story that was actually unusual. If you want the odds of making hands rather than running into them, poker hand probabilities owns that, and the odds chart has the printable tables.
AA vs KK: the exact numbers
The preflop split, and why the suits move it
Aces win 81.1% of runouts when all four suits are different, 81.7% when one king shares a suit with an ace, and 82.4% when both do. Chopped pots account for another 0.4 to 0.5%, so the aces’ equity, which is what you actually get paid on, runs 81.3% to 82.6%.
Suits move it for one reason: a king that shares a suit with an ace can no longer make a flush the aces do not also make. Take both of those escapes away and the kings lose about 1.4 points. It is a small effect and it is the only reason one number does not cover the matchup.
How often two players hold aces and kings at the same table
Being dealt a specific pair is 220 to 1 against, or about once in 221 hands. That part is the same for everyone.
The collision is the interesting number, and it is always stated from one player’s seat. Hold kings at a nine-handed table and one of the other eight has aces about 1 time in 26. Hold queens and about 1 time in 13 somebody has kings or aces, which is why queens are the hand that gets people into trouble. Hold aces yourself and another player has the other two aces about 1 time in 154. All three are counted before anyone acts, so they describe how often the situation arises, not how likely it is once a tight player has raised four times.
Why it always goes in
Because both hands are supposed to. Kings are the second best starting hand in hold’em, and a player who folds them whenever a stronger range might be out there loses far more to the folds than they save on the coolers. Neither seat made a mistake. The money went in with both ranges doing the right thing, and a bit less than one time in five the kings hold and nobody posts about it.
The other famous races
What are the odds of a pair against two overcards?
The pair is a small favourite, roughly 53% to 57% against ace-king offsuit. Deuces run about 53%, pocket tens about 57%, pocket queens about 57%. Suits matter more than people expect: against ace-king suited the queens drop to about 54%, because the suited hand adds flushes. This is the matchup people call a race, and the pair is a favourite in all of them.
AK against QQ: who is ahead?
Queens, at about 57% against ace-king offsuit and about 54% against ace-king suited. That narrow gap is why the matchup is the standard preflop argument in tournaments: the queens are ahead, and never by enough for anybody to feel good about it.
Kings against ace-king is a different animal and gets mistaken for the same race. Kings are about 70% there, because they hold one of the cards ace-king needs. That is domination rather than a race.
How often do aces win against a random hand?
About 85% against one random hand, which is the figure books quote for aces against any two cards. It falls as opponents pile in, because more players means more ways for somebody to make two pair or a straight, which is the reason a raise with aces is worth more than a limp.
How often is a set beaten by a bigger set?
About 1 time in 96, once you condition it properly: two players each holding a pocket pair, both flopping a set. You flop a set with a pocket pair around 11% of the time on its own. Set over set feels legendary because both players are delighted and only one of them is right, not because the number is astronomical.
An overpair against a flush draw: who is the favourite?
The overpair, at about 65%, if the money goes in on the flop and the draw has nine clean outs and nothing else. Nine outs complete by the river 35% of the time, and 19% on any single card. Whether the draw should be putting money in at that number is a question about the price it is being offered, which is a different page’s job.
How rare was your cooler?
| Matchup | Favourite | Favourite’s share | How often the situation turns up |
|---|---|---|---|
| Aces against kings, all in preflop | Aces | 81 to 83% | About 1 in 26 of the times you are dealt kings |
| Aces against one random hand | Aces | About 85% | You are dealt aces about 1 in 221 hands |
| Queens against ace-king offsuit | Queens | About 57% | About 1 in 13 of the times you are dealt queens, somebody has kings or aces |
| Kings against ace-king | Kings | About 70% | Domination, not a race |
| Two pocket pairs, both flop a set | The bigger set | Almost everything | About 1 in 96, given both players hold pairs |
| Overpair against a nine-out flush draw, all in on the flop | The overpair | About 65% | Depends on the board, not on the deal |
Every percentage in that table is for the money going in before the flop, except the last row, which is all in on the flop. The frequency column counts how often the situation arises from one player’s seat, before anybody has acted.
What these odds change about how you play
Preflop, almost nothing. You do not fold kings because aces exist somewhere in the room. The 1-in-26 number is a base rate counted before anybody acts, so it is a description of how often the spot occurs, not a licence to call once a specific opponent has taken a line that only makes sense with one hand. That call is a read, and reads are made at the table.
What they should change is what happens next. Losing a pot you were 82% to win is not evidence that you did something wrong, and treating it as evidence is where the real money goes. It is the same statistical shape as a losing week, which variance covers properly, and it is the moment tilt does its damage. For the odds of making rare hands rather than colliding with them, royal flush odds sits at the far end of that scale.
Common questions
Should you ever fold KK preflop?
Almost never, and books put the bar very high. One says that in a cash game it would never fold kings before the flop, and that folding is a little easier in a tournament. The exception they describe is always the same shape: a player you know to be tight takes a line so strong that it only makes sense with aces, usually a fourth raise or an all-in over the top of several. One older book turns that into a rule of thumb, that the fourth raise means aces, which is a read on a particular kind of opponent rather than a fact about the game. Against ordinary aggression, folding kings costs far more than paying off the occasional cooler.
What is a cooler in poker?
A hand where two players both have something strong, both play it correctly, and one of them still loses everything. Aces against kings is the standard example: nobody misplayed anything, the money went in, and the deck decided. The cooler entry has the full definition. The reason the word exists is to separate these hands from mistakes, because the two feel identical and only one is worth studying.
What are the odds of being dealt pocket aces twice in a row?
About 1 in 48,841. Aces come once in every 221 hands, and the deck has no memory, so two in a row is 221 times 221. It feels rarer than it is only because you are counting from your own seat. Across every table running right now, somebody is doing it as you read this.