Three-Betting in Omaha

You cannot blow anybody off a hand

Three-betting in hold'em earns money two ways: sometimes you are ahead, and sometimes everybody folds. Here is what the second half looks like in Omaha.

Take A♥A♦K♥K♦ — 70% against the button's whole opening range. Now sample 500 hands out of that range (126165 combinations) and run each of them against those aces for 20000 simulations:

• 51.2% of the button's range still holds 30% or more against the best hand you can have.
• The worst hand in the sample still had 8.9%.

Same measurement in hold'em, hero holding A♠A♥, 500 random hands at 10000 simulations each: 0 of them reached 30%, and the worst had 5.5%. To be exact rather than dramatic: one holding does clear the line — the remaining pair of aces, 1 combination out of the 1225 two-card hands left in the deck.

Half the range has a real hand against your best hand. Folding equity, the thing you have been buying with three-bets for years, is mostly not for sale here.

A three-bet earns its money from equity and pot-building, not from making people fold.

Six-handed at 100 big blinds against opponents playing a normal game — the baseline this course is built on.

Against a genuinely tight opponent who folds a lot, fold equity is real money and you should take it. Two cautions: that is an exploit, not the baseline, so it should change when he does; and you cannot bank it twice — you may not assume both "he folds often" and "when he calls I am still well ahead", because the hands he keeps are exactly the ones that had equity to begin with.

Strong and connected

Against the same button opening range, over 300000 simulations each:

• A♥A♦K♥K♦: 70%.
• A♥A♦8♣3♠: 60.4%.

9.6 points, and they are worth more inside a three-bet than they are in an opening pot. A three-bet builds a pot and shortens the stack behind it; the deeper you are, the more freedom you have to play a small pot and duck trouble, and the shallower you are, the more directly raw hand strength converts into chips. When you choose to build, you want the version of the hand that keeps being good after the flop — the second pair, the connected cards, the suits that reach the nuts.

Three-bet the hands that are both strong and connected.

At 100 big blinds, heads-up, when you intend to keep putting money in after the flop.

Our reference range still three-bets A♥A♦8♣3♠ from the big blind. Its 60.4% is a large preflop edge on its own, and passing it up costs more than the structure penalty does. "Strong and connected" is a ranking rule — when choosing what to three-bet, prefer the connected version — not an entry requirement. Run it as a gate and you turn a hand that is ten points ahead into a flat call.

How often, and why we phrase it carefully

In the one defending spot we have a table for — big blind against a button open — this reference range three-bets 10% and defends 36.1% in total. So roughly three of every four defended hands are calls.

You will often hear that three-betting in Omaha should be rarer than hold'em instinct suggests. This course passes that on as reasoning, not as a measurement, and the distinction matters enough to spell out: we have not measured a hold'em three-betting frequency to compare against, so we are not going to quote one. What we did measure is the previous page — 51.2% of the opener's range keeps 30% or more against the best hand in the deck. Remove fold equity from the reasons to three-bet and only value remains; hands with enough value to build a pot are a small set; the frequency follows from that, not from a chart.

When you defend, call more often than you three-bet; save the raise for real value.

Big blind against a button open at 100 big blinds — the one spot our table actually covers.

Calling keeps the pot small and the stack-to-pot ratio high, and a high ratio is precisely where bare aces are hardest to play (C5). So "prefer calling" is right for speculative hands and backwards for a pair of aces: the same hand class that most wants a small pot is the one that least wants one. A frequency rule that does not say which hands it applies to will get one of the two groups wrong every time.

The 10% and 36.1% are properties of our own table, counted the same way as the opening frequencies in C7 (combination-weighted, a mixed entry counting as half a hand). The comparison with hold'em on this page is an argument from the measurement on the previous page, and is labelled as such rather than dressed up as data.

What to take away

• Fold equity is mostly gone. 51.2% of an opening range keeps 30%+ against the best hand in the game; the same sample in hold'em produced 0.
• So a three-bet is a value-and-pot-building play. If a hand does not want a bigger pot with the money behind that a three-bet leaves, do not three-bet it.
• Prefer connected strength — worth 9.6 points on the same pair of aces — but treat that as a ranking, not a gate.
• The frequency that follows is small: 10% in the one spot we have a table for.

The three-bet you just made will often get called rather than folded to. What happens next — the pot size you have created and the money left behind it — is the whole of the next lesson.