How Engines Changed Human Chess

Before engines, opening preparation meant studying books and magazines. Now it means running Stockfish at depth 40 and finding novelties that no human has ever seen. The result: opening theory has advanced more in the last twenty years than in the previous hundred. Lines that were considered "equalising" in 2000 have been refuted; positions that were dismissed as "too complex" have been solved.

See also: How AlphaZero Changed Chess for the neural-network revolution.

Engines taught humans that evaluation is more nuanced than "count the pawns." The ChessBhumi analysis engine evaluates positions based on hundreds of factors: king safety, piece activity, pawn structure, mobility, and more. Humans who study engine evaluations develop a more sophisticated sense of what makes a position good or bad, beyond simple material counting.

Endgame tablebases (Syzygy, Lomonosov) have solved every position with seven or fewer pieces on the board. This means that every endgame with seven or fewer pieces has a known result, win, draw, or loss, with perfect play from both sides. This knowledge has changed how GMs approach endgames: they can trade into tablebase positions knowing the exact result.

The fear was that engines would homogenise chess, everyone would play the same "engine-approved" moves. Instead, engines have expanded the range of acceptable moves by showing that many "anti-positional" moves (sacrifices, pawn pushes, piece retreats) are actually objectively good. AlphaZero's games, in particular, inspired a wave of creative play at the top level.

The engine is a tool, not a crutch. Use game review to identify your mistakes, study engine suggestions to understand the ideas behind them, and use engine evaluation to calibrate your own positional judgment. But don't just mimic engine moves, understand why they work. That understanding is what makes you a better player.