← The Immortal Game

The Slow Lake Academy · The annotated game · ages 15-18

Also at: Chapter book

The Immortal Game, annotated

London, June 1851, a casual game during the first international tournament. Adolf Anderssen (White) v. Lionel Kieseritzky (Black). King's Gambit Accepted, Bishop's Gambit; Bryan Countergambit. 1–0, 23 moves as usually published.

Document 1. What we know, and how

Every move below is the standard published score, replayed by code before this page went up: each move is legal, each check is a check, the forced replies are forced, and the last move is mate. The ClassicsForge bank on this site holds the score and is the authority if any copy disagrees.

Two things are less solid than the moves look. First, Kieseritzky himself indicated that the game ended before checkmate; the famous final combination may have been worked out afterwards rather than played. Second, published versions disagree about the order of some moves, and some have plain errors. The name came later too: Ernst Falkbeer called it "the Immortal Game" in 1855. Kieseritzky published the game himself, in the Paris magazine La Régence, in July 1851, which is the earliest source. What you are reading is the canonical version of a real casual game, polished by a century and a half of retelling.

Document 2. The score, with notes

1. e4 e5 2. f4. The King's Gambit. White offers a pawn to open the f-file and deflect Black's e-pawn from the centre. It was the main line of attacking chess in 1851.

2... exf4 3. Bc4 Qh4+ 4. Kf1. The check forces the king to move and gives up castling for good. Ke2 and g3 were legal, and both are worse. White accepts an exposed king in exchange for time.

4... b5. The Bryan Countergambit. Black returns the pawn to pull the bishop off the a2–g8 diagonal, away from f7.

5. Bxb5 Nf6 6. Nf3 Qh6 7. d3 Nh5. Black's queen and knight are the only black pieces in play, and both are on the edge of the board. Material is level.

8. Nh4 Qg5 9. Nf5 c6. The knight on f5 eyes g7 and will stay there until move 21. Black attacks the bishop.

10. g4! A pawn hits the h5 knight that shelters the queen. The pattern of the whole middlegame starts here: White's pawns take squares from the black queen one at a time.

10... Nf6 11. Rg1! White leaves the b5 bishop to be taken and plays a rook move that supports g4 and later g5. 11... cxb5. White is now a bishop down.

12. h4 Qg6 13. h5 Qg5 14. Qf3. Each pawn move costs the queen another square. 14.Qf3 threatens 15.Bxf4, which would trap her.

14... Ng8. Black's knight retreats to its starting square to give the queen f6. In the language of development, Black has gone backwards.

15. Bxf4 Qf6 16. Nc3 Bc5. Black finally develops a bishop, and it attacks the g1 rook.

17. Nd5. White ignores the rook, attacks the queen and eyes c7.

17... Qxb2. The first grab. The queen leaves the defence entirely for a pawn on the far wing, with Qxa1+ to follow.

18. Bd6!? White offers both rooks. The move attacks the c5 bishop and, more importantly, takes e7 and f8 from the black king.

18... Bxg1. Steinitz argued in 1879 that 18...Qxa1+ first was better, while White would still have had chances. Note it, because it means the most famous attacking game in history may not have been forced from here.

19. e5! The quiet key move. The pawn blocks the long diagonal from b2 to g7, so the black queen can no longer return to defend her king. It also takes f6 from the knight on g8. This is interference: a piece placed across the line between a defender and what it defends.

19... Qxa1+ 20. Ke2 Na6. Black has won both rooks and is ahead by roughly a rook and a bishop. The knight heads for c7.

21. Nxg7+ Kd8. Forced: d8 is the only legal square.

22. Qf6+!! The queen sacrifice. It deflects the g8 knight, the one piece guarding e7. If 22...Ne7, then 23.Qxe7# (or 23.Bxe7#).

22... Nxf6 23. Be7#. Check the squares: c7 is covered by the d5 knight, e8 by the g7 knight, c8 and d7 are blocked by Black's own pieces, and the e7 bishop is protected by the d5 knight. Mate by a bishop and two knights. Black's rooks on a8 and h8 and bishop on c8 have never moved.

Document 3. Two columns

Keep two counts through the game. Material, by the usual values (pawn 1, knight and bishop 3, rook 5, queen 9). And pieces in play: pieces that can reach the decisive area within the next two or three moves.

After 11...cxb5, White is three points down and has five pieces in play against Black's two. After 19...Qxa1+, White is thirteen points down and has five pieces in play against Black's none that matter: the queen is cut off on a1, the bishop is on g1, the a6 knight is a move from relevance, and four pieces have never moved.

The usual piece values assume a piece can join the fight when it is needed. In this game that assumption fails for almost all of Black's material. The values are averages over typical positions, and this position is not typical.

Document 4. Interference and deflection

The finish uses two related ideas that are easy to blur.

  • Interference (19.e5) puts a piece across a line, so that a defender can no longer see its charge. The defender is not moved. It is cut off.
  • Deflection (22.Qf6+) forces a defender to move, so that it no longer guards its charge. The defender is not cut off. It is pulled away.

Both attack the relationship between a defender and what it defends, rather than the defender itself. That is the useful generalisation: when a position will not yield to direct attack, look at what each defending piece is responsible for, and ask how to separate it from that responsibility.

Document 5. The objections

First, it was a casual game, and Kieseritzky was playing for entertainment. His greedy captures would be judged harshly in a serious game.

Second, Black's 18th move was probably not the best, and with better defence the attack may not have been decisive.

Third, the ending may have been composed rather than played. If so, the published game is partly a puzzle attached to a real game, which is a different kind of object. It still teaches the same lesson, but it is evidence about what someone found afterwards rather than about what happened at the board. We do not know who first wrote the mate down, and it is worth saying so rather than guessing.

None of this makes the game worthless. It makes it a demonstration rather than proof. It shows what activity can do against material when the defender cooperates. Whether activity beats material against best defence is a different question, and the honest answer is "sometimes".

Document 6. Romantic chess and its end

The Immortal Game is the emblem of what is now called the Romantic era of chess, roughly the first two thirds of the nineteenth century. Players opened with gambits, attacked the king as early as possible, and treated a refused sacrifice as faintly cowardly. Defence was underrated, partly because defensive technique was not yet well understood.

The era ended largely through the work of Wilhelm Steinitz, the same player who later criticised Black's eighteenth move here. Steinitz argued that a sound attack has to be justified by an advantage in the position, and that a well-placed defence can absorb and refute an attack that is not. His ideas became the basis of what is called positional play.

The two views are not as opposed as they sound. The Immortal Game is a good example of Steinitz's point as well as the Romantics'. White's attack works because Black's pieces are badly placed. That is a positional advantage, earned move by move with pawn moves and tempo gains, long before any sacrifice. Anderssen's sacrifices cash in an advantage he already had. Read that way, the game is less a triumph of daring over material than a triumph of position over material, with daring as the method of collection.

Document 7. The gambits at the start

Both sides open with a gift. White's 2.f4 offers a pawn to pull Black's e-pawn away and open the f-file. Black's 4...b5 offers one back to pull White's bishop off the diagonal toward f7. Each gambit is a bet that time and lines are worth more than a pawn. By move five, material is level again and both sides have spent a little time. What separates them afterwards is not the gambits but what each side does with the next fifteen moves: White uses them to develop and take squares, Black uses them to move the queen.

Document 8. Your position

Set up the position after 18...Bxg1. Before you look at 19.e5, list every white move that gives check or captures. Then find the move that does neither and is stronger than all of them. What did you have to stop looking at to find it?

Then set up the final position and count, for each black piece, how many moves it would need to reach a square next to its own king. Add them up. That number, not the material count, is the real measure of Black's position at the end. Do the same for White's pieces and compare. The gap between the two totals is a crude but honest way to put a number on activity.

The Honest Page

Which parts of this story are true, stretched or made up?

True

  • Every move in this story is the game as it is usually published: Adolf Anderssen against Lionel Kieseritzky, twenty-three moves, mate.
  • It was a casual game, played in London in 1851 during a break in the first international chess tournament. Source
  • At the end, White had given up a bishop, both rooks and the queen. Black's two rooks and queen's bishop had never moved.

Stretched

  • The story plays the game all the way to mate. What we simplified: Kieseritzky himself said the game ended before mate, and the last few moves may never have been played on the board. Versions also disagree about the order of some moves. Source
  • It is called the Immortal Game. What we simplified: Nobody called it that in 1851. Ernst Falkbeer gave it the name in 1855. Source

Made up

  • Captain Castle's academy, the re-staging and chess pieces who talk are made up.