Chess terms: Computer chess and analysis
30 terms about computer chess, each with a plain definition and links to related terms.
Accuracy is a percentage produced by analysis software that summarizes how closely a player's moves matched the engine's preferred choices over a game, with higher numbers meaning fewer costly errors.
Alpha-beta pruning is a search technique that skips branches of the move tree that cannot change the minimax result, letting engines look much deeper using the same amount of computing time.
Average centipawn loss, or ACPL, is the mean amount of evaluation a player gave up per move compared with the engine's best choice, measured in hundredths of a pawn, so lower figures mean cleaner play.
The best move is the move an engine rates highest in a given position after searching, usually shown first in its list of candidate lines and marked with the strongest evaluation.
A blunder is a very serious error, such as leaving a piece undefended or missing a decisive tactic, that swings the evaluation so far that a good position turns bad or a bad one becomes lost.
A blunder check is a quick test, made by a person before moving or by software afterward, that looks for a move which hangs material, allows a tactic or loses the game outright.
A brilliant move is a label some analysis tools give to a strong, hard-to-find move, often a sacrifice, that the engine confirms keeps or improves the position despite looking risky.
A centipawn is one hundredth of a pawn, the small unit engines use to report evaluations, so a score of 150 centipawns means the engine thinks one side is about one and a half pawns better.
A chess engine is a computer program that searches positions and evaluates moves, usually playing and analyzing far stronger than any human, and it is used for analysis, preparation and play against a machine.
A chess database is a searchable collection of games, stored in a standard format, that lets a player look up positions, study opponents, and see how often moves have been played and what results they scored.
An endgame tablebase is a precomputed database that gives the exact result and best moves for every position with a small number of pieces, so engines can play those endgames perfectly.
Engine analysis is the use of a chess program to search a position and report which moves are strongest, along with a numeric evaluation and the lines it expects both sides to follow.
An evaluation is a numerical judgment of who stands better in a position, usually given in pawns from White’s point of view, where positive favors White and negative favors Black.
The evaluation bar is a vertical or horizontal gauge beside the board that shows, with a bar of black and white, which side the engine thinks is better and by roughly how much.
Fair play is the rule that every game must be decided by the players' own thinking, which means no engine help, no advice from others during play and no tricks that distort a result.
A game review is an automatic engine pass over a finished game that marks each move, shows how the evaluation changed and points out the turning points, mistakes and missed chances.
An inaccuracy is the mildest error label in game review, given to a move that is playable but gives up a small amount of advantage compared with the engine's better alternative.
Minimax is a decision method in which one side tries to maximize the evaluation while the opponent tries to minimize it, forming the basis of how traditional chess engines choose moves.
A missed win is a game review label for a move that fails to take a clearly winning continuation, so a decisive advantage shrinks to something much smaller or disappears entirely.
A mistake is a game review label for a move that clearly worsens a position, costing a meaningful share of the advantage, though it is less severe than a blunder that throws the game away.
A neural network engine uses a trained neural network to evaluate positions, and NNUE, the efficiently updatable neural network added to Stockfish, combines such an evaluation with fast traditional search.
An opening book is a database of opening moves and positions that an engine or program consults to play known theory quickly, instead of calculating each early move from scratch.
An opening explorer is a tool that shows, for any position reached from the start, which moves players have chosen in past games and how often each move led to a win, draw or loss.
The principal variation is the line of best play for both sides that an engine has found, shown as a sequence of moves and tied to the evaluation it reports.
A puzzle rating is a number that estimates how hard a tactical puzzle you can reliably solve, and it rises and falls as you solve or fail puzzles of known difficulty, much like a game rating.
A puzzle theme is a label for the tactical idea a puzzle teaches, such as a fork, a pin or a back rank mate, so that training can be grouped by pattern and weaknesses can be targeted directly.
Search depth is how many moves ahead an engine has looked in a position, counted in half-moves called plies, and deeper searches generally give more reliable evaluations and moves.
Stockfish is a free, open-source chess engine developed by a worldwide community, widely regarded as one of the strongest engines available and used by many analysis websites and apps.
Syzygy tablebases are a compact set of endgame tablebases created by Ronald de Man that cover endgames up to seven pieces and are widely used by modern chess engines.
Winning chances are an estimate of how likely each side is to win from a position, often converted from the engine's evaluation into a percentage so that the same advantage reads differently at different levels.
Other topics
Learn the ideas on the board, not only the words
A first lesson starts with your games: George reviews them and builds a study plan around what he finds.