The final table and heads-up: closing out a tournament

Hand values are relative to how many players can wake up with something better. A9o is a marginal hand at a 9-handed table — and a monster 3-handed. As the final table loses players, every range you have should widen: opens, defends, 3-bets, jams.

A useful mental model: your position is improving even when your seat isn't. At 9-handed, an under-the-gun open shows a strong range because seven players are still to act. At 4-handed, the same first-to-act seat is effectively a cutoff. Players who keep their full-ring standards at a short table blind themselves away in the most expensive seats of their tournament life.

Final-table pay jumps are steep — the difference between 6th and 5th can exceed everything you've laddered so far. ICM makes two opposite plays correct at once:

• With a fragile medium stack: folding genuinely strong hands to all-in pressure is often right when a shorter stack is about to bust. Let the 3bb stack blind out before you take a flip — that's not weakness, it's arithmetic.
• With the big stack: you should be raising relentlessly *because* everyone else is correctly waiting. The medium stacks are mathematically barred from playing back at you without premium hands. This is the most profitable bullying licence in poker.

The sin isn't folding or shoving — it's doing either without knowing why. Check the stacks before every decision: who busts next if things go wrong, and what is that worth to you in real money?

Three-handed play combines relentless blind pressure with the most awkward ICM of the whole event. You post a blind in two of every three hands — sitting back is bleeding at maximum speed.

Working principles:

• The button is gold. You have position on both opponents and only two hands can be out against you. Opening around half your hands is a starting point, more if they're folding.
• Stack ranking dictates aggression. The chip leader attacks the second stack hardest (the second stack loses the most by busting third). The short stack often has *more* licence to jam than the middle stack has to call.
• Pair-versus-position battles decide it. Most three-handed pots are unpaired-versus-unpaired; the player who realises equity in position wins the war of small pots.

Once you're heads-up, second place is locked up. You're now playing for the *difference* between first and second — and that difference is pure chips. The survival tax that governed the final table is gone: heads-up is essentially chip-EV poker again, and the tight habits that laddered you here become leaks the moment the third player busts.

What changes concretely:

• Any pair, any ace is a strong hand. Against one random hand, A2o is roughly a 55% favourite and most pairs are 70%+ to be best preflop. Hands you'd muck 9-handed are now your value range.
• The button should open most hands — with position and the small blind already posted, folding much more than a third of your hands is surrendering the match.
• Aggression compounds. The player who wins the small uncontested pots can lose the big showdowns and still climb. Passive heads-up play loses even with better cards.

Most players reach their first final tables exhausted and play not to lose. The field's tendency at every final table is too tight, too passive, too eager to ladder — which means the correct adjustment is usually to be the one applying pressure, not absorbing it.

That doesn't mean recklessness; it means knowing which force governs the current decision. Covered by three stacks with a pay jump looming? ICM — tighten and let others bust. Chip leader, or heads-up? Chips — attack. The trophy goes to players who can tell the difference in real time, hand after hand. Practise both gears here, where final tables cost nothing.