Nut Awareness

The nuts is not a feeling

Board T♥9♦8♣4♠2♥. What is the best possible hand? Not "a straight". The best possible hand is Q♠ plus J♠ — or any other queen with any other jack. Because you must use 2 cards from your hand and no more, the absolute nuts on any board is always one specific two-card combination, and here there are exactly 16 of them (four queens times four jacks).

Rule of the game: every showdown hand uses exactly 2 hole cards and exactly 3 of the 5 board cards. That is why the nuts is a two-card combination — on a flush board just as much as on a straight board.

This is the definition of Omaha, not a heuristic. Our rule layer computes the nut value by enumerating every two-card combination against the board (the postflop rule spec), which is also how it counts the 16 above.

Now count how often somebody has it. Your opponent holds four cards, which is 6 different two-card combinations — six chances to contain one of those queen-jack pairs. Enumerating every possible opponent hand on that board:

• Omaha, opponent holds four cards: 7.73% of hands are the absolute nuts.
• Hold'em, opponent holds two cards: 1.48% — a single combination and no second chance.

That is 5.2 times as often. Nothing about the strength of the nuts changed; what changed is how crowded it is. Every warning in this camp comes out of that one ratio.

What second best actually costs

River K♥8♥3♥2♠7♦. Three hearts, so the absolute nuts is the ace-queen of hearts. Here is what three different flushes are worth against opponents who are actually putting money in — we deal them hands that have two pair or better on this board and count how often you win the pot:

• A♥Q♥2♣5♦ — the nuts, rank 1: 100% against 3, 4 or 5 players.
• Q♥T♥2♣5♦ — rank 11: 83.3% / 75.2% / 65.4%.
• 6♥4♥2♣5♦ — rank 41: 46.5% / 27.8% / 14.7%.

In a multiway pot the hand that collapses is not "second best" — it is the bottom of a category. Adding players costs the eleventh-best flush a slice; it costs the small flush most of what it had.

Measured on the river, on a complete five-card board, against opponents restricted to hands that would put a stack in. Our rule layer refuses to give a verdict for multiway flops and turns because this project has never calibrated that case — two more cards are coming and everybody's share is lower for reasons that have nothing to do with the hand.

The opposite mistake is just as expensive. "If it is not the nuts, fold" throws away far too much: with 4 opponents who each hold trips or better, the chance that somebody really is holding the absolute nuts is only 29.3% (250 random river boards, all 178,365 opponent hands enumerated per board, opponents treated as independent of each other). Roughly seven times out of ten, nobody has it.

The qualitative direction — tighten multiway, aim for "nut potential" rather than "the nuts" — is public. GTO Wizard, "10 Tips for Multiway Pots" <https://blog.gtowizard.com/10-tips-multiway-pots-in-poker/> That article is about hold'em, not Omaha, and it gives no numbers. Every percentage above is ours, 40,000 restricted deals per cell.

Rank order is not strength

Same river K♥8♥3♥2♠7♦. A♥T♥2♣5♦ is not the nuts — it is rank 3, two whole places off. It also wins 100% against three opponents and 100% against five. Rank 1 and rank 2 do exactly the same. On this board the top of the list is a run of "ace of hearts plus another heart" hands that are, in practice, the same hand.

Then the list falls off a shelf: rank 11 is 83.3%, rank 41 is 46.5%.

Do not ask "how far from the nuts am I". Ask "which specific two-card combinations beat me, and would somebody be putting this much money in with them".

Always, but it bites hardest on flush boards, where the ranking list is long and its top is nearly flat.

A rule written on rank order gets this board exactly backwards. "Play rank 2 or better" folds rank 3, which never loses, and still does not protect you from rank 11, which wins 83.3% and is a fine call. The distance between ranks is even; the distance between hand strengths is not.

This is why the tool we give you in this camp is a number you can compute — a share of pots won against opponents who are betting — and never a position in a list.

What to take away

• The nuts is a two-card combination, and it is crowded. An opponent holds it 7.73% of the time on our test board, against 1.48% in hold'em — 5.2 times as often.
• Second best is fine; the bottom of a category is not. Rank 3 wins 100%, rank 11 wins 83.3%, rank 41 wins 46.5% against three players.
• "Not the nuts, so fold" is a leak. With 4 opponents holding trips or better, somebody actually has the nuts only 29.3% of the time.
• Read the board before you read your hand. Name the exact two cards that make the nuts. Then ask who would be here with them.

Everything in this camp is a rule of this course, not an equilibrium solution. Each threshold is a line we drew because our own engine showed it separates situations that behave differently; none of it comes from a solver, and we never pretend otherwise.

Next: how many cards actually improve you, and how many of those improvements are worth having.