Why it matters

Minimax is game theory foundation. Understanding shapes game AI history.

Advertisement

The architecture

Max nodes: pick child with max value.

Min nodes: pick child with min.

Leaf: evaluation function.

Minimax evaluationMax nodepick highestMin nodepick lowestLeafstatic evalAssumes optimal opponent; risk-averse; alpha-beta prunes to same result
Minimax.
Advertisement

How it works end to end

Backup values up tree.

Assumes perfect play by opponent.

Basic exponential in branching factor.

Alpha-beta pruning essential.