Chess Games Worth Knowing By Heart
Memorising whole games is not the same as memorising openings. The point is not to reproduce the moves at the board — it is that a game you know deeply becomes a reference point: you recognise the pattern when a similar structure appears in your own games. These are the games that come up again and again in teaching, chosen because each one demonstrates a single idea cleanly enough to be worth the effort.
Start Here: Three Classics
- Morphy vs. Duke of Brunswick and Count Isouard, Paris 1858 — the "Opera Game". Seventeen moves, played in a theatre box, and still the single best demonstration of rapid development and open lines ever recorded. If you memorise one game, make it this one: every move has an obvious purpose, which is exactly what makes it teachable.
- Anderssen vs. Kieseritzky, London 1851 — the "Immortal Game". Anderssen gives up both rooks and the queen and mates with minor pieces. Romantic-era chess at its most extreme, and a useful corrective to the idea that material is what matters.
- Anderssen vs. Dufresne, Berlin 1852 — the "Evergreen Game". The companion piece to the Immortal, with a famous queen sacrifice and a mating net that is worth being able to see coming.
Positional Understanding
- Rotlewi vs. Rubinstein, Lodz 1907 — often called Rubinstein’s Immortal. The build-up is pure positional play; the finish is a combination that only works because every piece was already where it needed to be. The lesson is that tactics flow from position.
- Steinitz vs. von Bardeleben, Hastings 1895 — a long forcing sequence where the defender simply left the room rather than face the mate. Excellent for training calculation of long forced lines.
- Botvinnik vs. Capablanca, AVRO 1938 — famous for a piece sacrifice that creates unstoppable connected passed pawns. Worth knowing for the endgame judgement as much as the combination.
- Karpov vs. Kasparov, Game 16, World Championship 1985 — the "octopus knight" on d3. The best single illustration of what a permanently outposted piece is actually worth.
Modern Games
- Donald Byrne vs. Bobby Fischer, New York 1956 — the "Game of the Century", played when Fischer was 13. The queen sacrifice on move 17 and the windmill that follows are the reason this game is in every anthology.
- Fischer vs. Spassky, Game 6, Reykjavik 1972 — Fischer plays 1.c4 for essentially the first time in his career and produces a game so clean that Spassky applauded it. A model of quiet positional accumulation.
- Kasparov vs. Topalov, Wijk aan Zee 1999 — probably the most celebrated attacking game of the computer era, featuring a rook sacrifice and a king hunt that drives Black’s king from g8 across the entire board.
- Short vs. Timman, Tilburg 1991 — the king walk. Short marches his own king up the board to deliver mate in the middlegame. Short, memorable, and it permanently changes how you think about king safety.
How to Actually Memorise a Game
Rote repetition of move sequences does not work well and does not transfer to your own play. What works better:
- Understand before you memorise. Play through the game slowly with annotations first, until you can explain why each move was played. Moves you understand stick; moves you do not will not.
- Chunk it by phase. Learn the opening, the critical middlegame moment, and the finish as three separate units rather than one thirty-move string.
- Guess the move. Cover the game score and try to find each move yourself before revealing it. Where you go wrong is where the actual learning is.
- Replay from memory on a real board. Physically setting up and playing it through is a much stronger memory cue than clicking through on a screen.
- Space the repetitions. Revisit a game after a day, a week, then a month. Three spaced passes beat ten in one sitting.
Ten games known thoroughly are worth far more than a hundred played through once. Start with the Opera Game, and only move on when you can play it from memory and explain every move.