Heads-up versus Multiway

Adding players does not tax everyone equally

River K♥8♥3♥2♠7♦. We deal opponents hands that have two pair or better here — hands that would actually be putting money in — and count each holding's win share — the share of the pot it takes down, ties split — over 40,000 deals per cell. Read across: one opponent, then two, three, four.

• Q♥A♥2♣2♦, rank 1: 100% … 100%. Nothing touches it.
• Q♥T♥A♠4♣, rank 11: 92.5% · 84.7% · 76.6% · 68.1%.
• J♥T♥A♠4♣, rank 18: 86.1% … 47.7%.
• 6♥4♥A♠9♣, rank 41: 72.5% · 49.9% · 31.5% · 17.7%.

Multiway does not shift your whole range down one notch. It deletes the bottom of each category and leaves the top almost untouched. Adjust by cutting the bottom, not by tightening uniformly.

River, complete board, opponents restricted to hands that would commit chips. Our rule layer refuses to give a multiway verdict on flops and turns at all, because this project has never calibrated that case.

"Not the nuts, so it is trash" is a sentence with a missing argument. Rank 41 — the fourth-worst flush on the board — wins 72.5% against one opponent who is betting. That is a value hand. The same four cards against four opponents win 17.7%. Nothing about the hand changed.

Multiway punishes being dominated, not being outnumbered

Two draws on 9♥8♦2♣, both against top set, and then we add a third player. All numbers below are full enumerations of every run-out, not samples.

• J♠T♥7♦6♣ (20/14) against top set: 54.4%. Add a middle set and it goes up to 57.2% across all 666 run-outs — the two sets are taking chips off each other.
• 7♣6♥5♦4♠ (13/3) against top set: 43.9%. Add a nut wrap and it collapses to 20.9%.

Before you tighten for the extra player, ask what the extra player is contesting. A hand that nobody is fighting you for does not get worse when the pot gets crowded.

Any street; these two are measured on the flop. The mechanism is about who is drawing to the same card as you, so it is strongest where the nuts is still moving.

A blanket "tighten by a category in multiway" pushes both hands above in the same direction, and their correct directions are opposite: one went from 54.4% to 57.2% when the extra player arrived, the other lost more than half of what it had. Rules that count seats instead of counting contested nuts cannot be right for both.

And one number to stop the overcorrection before it starts: with 4 opponents who each hold trips or better, somebody actually holds the absolute nuts only 29.3% of the time. That figure comes from 250 random river boards with all 178,365 opponent hands enumerated on each, then treating opponents as independent of one another — an approximation we make deliberately and are telling you about, because four opponents who all have trips cannot literally be dealt on most boards.

The adjustment goes through the nut requirement

You already have three stack-to-pot bands: mechanical below 1.5, planned up to 4, a real three-street game above it. The honest thing to say about them multiway is where they move — and we do not know.

This course has never calibrated multiway stack-to-pot bands, so it gives you no number for them. Inventing an adjustment factor is exactly the kind of made-up figure this product exists to avoid, and it would be undetectable — a plausible band is far more dangerous than an admitted blank. Measuring it needs multiway equity and a multiway betting tree; neither arrived with the full-street engine, and there is no date on them. Until then the bands are heads-up bands, full stop.

Make the multiway adjustment through the nut requirement — how good your hand has to be to keep going — and leave the ratio arithmetic alone. The arithmetic did not change when a player sat down; the competition for the nuts did.

Three or more players. With two, this whole lesson collapses into the previous ones.

Try it the other way and it breaks immediately: in the same 3-handed spot, one draw finished at 57.2% and the other at 20.9%. Any adjustment that reads only the number of players moves both the same way.

The ceiling from the pot-odds lesson is worth repeating here, because it puts a floor under how tight "tight" can be: facing a single bet you never need more than 33.3%, no matter how many people are in the pot. There is no sizing in this game that can demand more of a caller, so "multiway means fold almost everything" cannot be justified by the price.

What to take away

The sentence this lesson exists to kill is "multiway, only play the nuts". It is not a conservative version of the truth; it is a different claim, and our own numbers say it is wrong in both directions — it folds hands that win two pots in three, and it does not protect you from the hands that actually break you.

• Cut the bottom, not the whole range. Rank 41 goes 72.5% to 17.7% as opponents pile in; rank 11 only falls to 68.1%.
• Nobody has it as often as you think. With 4 opponents on trips or better, somebody holds the absolute nuts 29.3% of the time.
• Crowding punishes domination. 57.2% for the nut wrap, 20.9% for the dominated one, in the same three-handed spot.
• The price has a ceiling. Facing one bet, never more than 33.3%.

All of this is a rule of this course, not an equilibrium solution — and the one thing we refuse to give you here, the multiway stack-to-pot bands, is refused on purpose and dated.