In combinatorial game theory, misère quotients encode the essential structure of games played under misère rules (where the last player to move loses, inverting normal play). Rubinstein-Salzedo and Zhou prove that these quotients can have any finite cardinality — except three.

For partisan games (where the two players have different available moves), misère quotients can be any size: 1, 2, 4, 5, 6, and so on. But never 3. For impartial games (where both players see the same moves), the constraint is stricter: the cardinality must be even.

The prohibition on three isn’t an accident of the construction or a gap in the proof. It’s a structural impossibility — the algebraic constraints that define misère quotients exclude exactly this cardinality. Every other finite size can be realized by an explicit game.

What makes this striking is the asymmetry between partisan and impartial variants. Allowing the players to differ in their available moves opens up almost unlimited structural flexibility — any cardinality except 3. Forcing symmetry (impartial play) collapses the possibilities to even numbers only. The single degree of freedom (player asymmetry) has a dramatic effect on the algebraic space of games.

The deeper point: strategic flexibility doesn’t just expand the set of games you can play. It expands the set of algebraic structures those games can realize. But even maximal flexibility has exactly one forbidden value — and the reason it’s forbidden reveals a constraint that’s more fundamental than the games themselves.