Grading Your Aces

Not all aces are aces

This is the first hole every newcomer has, and it is expensive because it feels like knowledge. Put two different pairs of aces against the same opponent, J♣T♣9♠8♠, and enumerate every one of the 1086008 possible boards — no simulation, no error bar:

• A♥A♦K♥K♦: 61.5%.
• A♥A♦8♣3♠: 51.8%.

9.7 points apart, and both of them are "aces". The second one is a coin flip against a hand a lot of players would fold. Put them against each other — A♠A♥K♠K♥ versus A♦A♣8♦3♣ — and the good version is a 55.6% favourite over the bad one even though the aces cancel out entirely.

So "I have aces" is not a plan. The question is which aces, and the four dimensions from C4 answer it: what do the other two cards do, and are the suits working?

The two enumerated numbers come from test/fixtures/preflop-exact.js. The head-to-head is 1000000 simulations, standard error 0.05 points.

The ladder from bad aces to good aces

Against a random hand, one rung at a time:

• A♥A♦8♣3♠ — rainbow, two dead cards: 61.6%.
• A♠A♥8♠3♥ — the same ranks, double-suited: 67.4%. Suits alone.
• A♠A♥K♠Q♥ — double-suited with two big connected cards: 69.0%.
• A♥A♦K♥K♦ — double-suited, second pair, both suits ace-high: 70.7%.

And the ladder runs downwards too: A♠A♥J♠T♣ at 68.0% becomes A♠A♥J♠2♣ at 65.2% for the price of one dead card.

Good aces, in this course's language: aces plus at least one of these two — a second pair, or two connected big cards — with the suits doing something — ideally both suits ace-high, so the flushes you chase are the ones nobody beats. Bad aces: rainbow, plus two low cards that neither connect to the aces nor to each other. Everything between is a judgement call, and position decides it.

The reason this matters more here than the raw percentages suggest: good aces flop something extra — a set with a flush redraw, a nut flush draw over a made hand, a freeroll where you cannot lose and might win the whole pot. Bad aces flop one pair on a board that has passed them by. Same preflop label, different game after the flop.

How bad aces actually make money

Bad aces are not a folding hand — our reference ranges open A♥A♦8♣3♠ from every seat, under the gun included. They are a hand with one plan: get the money in while you are still the favourite, then have as little poker left to play as possible.

With aces, raise and re-raise to shrink the stack-to-pot ratio.

Heads-up, at 100 big blinds or shallower, when the re-raising actually gets you to a low SPR — and when your opponent is the kind who will put money in behind you.

Shrinking the SPR is a means, and it does not always arrive. Work through the four-bet line from C2: after four-betting you still have 125 chips behind, which is 62.5 big blinds. That is not a low SPR; that is three streets of postflop with a hand that flops one pair. Bare aces are 51.8% against a real connected hand over all 1086008 boards, and the low hand is not a soft spot either — 7♥6♥5♦4♦ holds 37.4% against A♥A♦K♥K♦, exactly enumerated. When the SPR is going to stay high, bad aces are one of the hardest hands in the game to play, and building a bigger pot with them makes it worse rather than better.

The practical version: raise them, be happy to get it in preflop, and if the pot stays medium and the flop misses you, be willing to play a small pot with one pair — which is exactly what C3 told you to do.

What to take away

• "Aces" is a range, not a hand. The best and worst versions are 9.7 points apart against the same opponent.
• Grade them with C4's four questions. Second pair or two connected big cards, plus suits that reach the nuts, is the difference between 70.7% and 61.6% against a random hand.
• Bad aces still open from everywhere — they just need a small pot or a finished one, never a medium one.
• Against a real hand, bare aces are 51.8%. If that number surprises you, it is the exact size of the leak this lesson is closing.

The next lesson runs the same four questions on hands with no pair at all: how high the connected cards run, where the gap sits, and when a suit outweighs a rank.