Rate moves by how likely they are to induce an opponent mistake

For each candidate move, plays it and measures how badly a human opponent is likely to go wrong in the resulting position. Moves are then scored on practical value: what the move costs you objectively, set against what it is likely to win you from the opponent's mistakes.

Usage

trappiness(fen, maia, ctx, top_n = 5L, movetime_ms = 800L)

Arguments

fen

Position to move from.

maia

A session from maia_session_start() representing the opponent.

ctx

A chess.js V8 context, used to apply moves.

top_n

How many of your own candidate moves to examine.

movetime_ms

Engine time per evaluation search.

Value

A data frame of move, cp (its objective value to you), cost (centipawns given up versus your best move), trap (the opponent's expected loss afterwards) and practical (trap minus cost), ordered by practical.