Use this search procedure to study or select actions in a defined game tree when legal moves, terminal values, depth, and evaluation rules are available.

Game tree

Required object input.

Maximizing root

Optional boolean input.

Evaluate a finite minimax tree while pruning branches that cannot affect the result. The tool evaluates the supplied inputs together and returns the named outputs below; it does not infer omitted operating conditions or change the units shown.^[1]

Search result

The resulting search result returned as an object.

• Search quality depends on the game model, evaluation function, horizon, move ordering, branching factor, stochastic assumptions, and computational budget. A returned move is optimal only within the searched model and bounds.
• Use finite inputs in the displayed units and preserve more precision than the final presentation requires. Independently verify safety-critical, financial, compliance, or production decisions.

Compare search depths and ordering strategies, verify small trees exhaustively, and test decisions across representative positions rather than relying on one example.

• Alpha-Beta UCT Calculator

  Rank Monte Carlo tree-search actions with UCT scores constrained by alpha and beta bounds.

• Varshni Bandgap Calculator

  Evaluate the Varshni temperature dependence of a semiconductor bandgap: Eg(T)=Eg0 - alpha·T²/(T+beta), using a logarithmic evaluation that avoids overflowing T². The coefficients are material-specific empirical fit parameters; extrapolation beyond their fitted temperature range can produce nonphysical results.

• Beam Search Visualizer

  Expand a scored tree while retaining only the best paths at each level.