Where does the values of alpha-beta search get updated?

Key idea: alpha and beta are updated during the search as nodes are evaluated; these updates drive pruning decisions.

• Initialize alpha and beta at the start of the search (commonly alpha = -infinity, beta = +infinity).

• When visiting a maximizing node, after evaluating a child with value v, update alpha = max(alpha, v).

• When visiting a minimizing node, after evaluating a child with value v, update beta = min(beta, v).

• After each update, check the pruning condition: if alpha >= beta then stop exploring remaining children of the current node (prune).

Worked summary:

• Updates happen while traversing the tree (along the search path), not only at initialization or at the end.

• These live updates allow the algorithm to prune branches early and avoid unnecessary evaluations.