The Right Exchange: Ding's 55.Rf2 and the Trade That Lost a Title

Part 1Game 14: one game left, scores level

In December 2024 Ding Liren defended the world title in Singapore against Gukesh Dommaraju, the 18-year-old challenger from India. After 13 games the score was 6.5–6.5. Game 14 was the last classical game; a draw would have sent the title to rapid tiebreaks the next day. Ding had White. ^[3] ✓ Verified

Gukesh won, 7.5–6.5, and became the youngest undisputed world champion. The game was level for nearly all of its 58 moves. What decided it was one exchange: a rook trade on move 55 that looked like the last step to a draw and turned out to be the first step to a lost king and pawn ending. ^[3][5] ✓ Verified

That makes it the cleanest modern lesson in exchanges. The defender's plan was to trade pieces until nothing was left to lose; the attacker's plan was to avoid the trades that made that easy. Both plans were sound. The game turned on the single trade where the usual rule (when defending, trade) stopped being true. For the same match seen through other strategies, see restriction in game 12 and the pawn sacrifice in game 11. ◐ Inferred

Part 2What is the right exchange in chess?

The right exchange is the trade that leaves you a better position than the one you had: you give up a piece for an equal one, and what remains suits your plan. The usual rules are good starting points (trade pieces when you are ahead, trade the opponent's active pieces when you defend, keep the piece that has no opposite number), but each rule fails somewhere. The only reliable test is the position after the trade.

Two kinds of exchange need special care. A trade that leads to a king and pawn ending can never be undone: there are no pieces left to repair a mistake, and the result is usually a forced win or a forced draw. And a trade that forces a second trade must be judged by where both trades end, not the first.

Part 3The opening: a reversed Grünfeld and an early wish to trade

Ding opened with 1.Nf3 and 2.g3, and Gukesh set up a Grünfeld with colours reversed: pawns on c5 and d5, then ...d4 to gain space. After 13...Bb6, which annotators called inaccurate, White had chances to press. Ding chose a quieter path with 14.b3, and by move 18 the knight on d5 was White's best piece. ^[4] ✓ Verified

The structure that followed was simple. After 19.cxb5 axb5 20.Nf4 exf4 21.Bxc6 Bxb2 22.Qxb2 the knights and the dark-squared bishops were gone, and by move 32 so were the queens and one pair of rooks. What remained was rook and light-squared bishop each, with Black's three kingside pawns against White's two. In such endings the defender's usual plan is to trade the remaining pawns and pieces. ^[1][2] ✓ Verified

Part 4The plans: trade everything, or keep the game alive

Ding's plan, from move 19, was to trade. Annotators read it as a decision to head for a draw and the tiebreaks: 19.cxb5, 20.Nf4, 26.a4, 29.b4 (a pawn given to force the queens off) and 31.Ra8 all removed material. Gukesh's plan was the opposite. With nothing to lose, Black kept the rooks and bishops on, pushed the kingside pawns to gain space, and waited for a slip. ^[4][2] ✓ Verified

The defender's counterplay depended on one thing: the white rook. As long as it blocked the f-pawn and guarded the fourth rank, Black's king could not break through and the g4 pawn could not advance safely. That rook was the piece Ding offered to trade on move 55. ◐ Inferred

Part 5The critical moments: three trade decisions

Trading when slightly better is not a mistake in a must-not-lose game. The cost is that the game now belongs to the other side: whoever has the only winning chances decides how long it goes on. Gukesh's 27...g6 and 28...Qb5 (avoiding a queen trade that White offered with 28.Qd4) made that clear. ^[4] ✓ Verified

From move 38 to move 54 the pattern repeats. Ding traded a pair of pawns with 45.f3, which Gozzoli called a very good defensive move, and kept the rook on the fourth rank. Gukesh walked the king to e5 and the bishop to e6. The clock, though, was moving one way: after 52.Ra4 Ding had about twelve minutes left, Gukesh more than an hour. ^[1][4] ✓ Verified

Part 6The turning point: one rook trade too many

At the first moment, I didn't see that I was winning. When I realized that I was winning, it was the best moment of my life.

The win is a matter of opposition: the kings stand face to face and the side to move must step aside. After 57...Kxd5, only 58...Ke5 wins against 58.Ke3; Stockfish rates the king moves to d6 and e6 as draws. Two exact moves by Black (56...Bd5 and 58...Ke5) were all the technique required. ✓ Verified

Part 7What Ding could have done: a fortress with pieces on

White's position on move 55 was a fortress: a position the stronger side cannot break into. The rook on f4 blocks the f-pawn and attacks g4; the bishop guards the long diagonal from a8 to h1, including e4, the square Black's bishop wants. Black's rook can roam the second rank, but there is nothing for it to win. ◐ Inferred

Rook exchanges work 99 percent of the time (except now when the bishop will also be exchanged as it is cornered on a8!)

The defender's rule is not 'never trade'. Ding's earlier trades were all sound. The rule is to check what the opponent can do the move after the trade. With the rooks off, the bishop on a8 had only three moves (b7, c6 and the capture on d5), and 56...Bd5 covered all of them. Putting the bishop on a8 two moves earlier, with 53.Ba8, was harmless while the rooks were on; it became fatal only once a rook trade made a bishop trade possible. ^[2] ✓ Verified

Part 8A different angle: Ding's trades were right, until one was not

The usual account of game 14 is that Ding, under pressure, blundered. That is true, but it hides the more useful lesson. Stockfish shows that Ding's drawing strategy worked: from move 19 to move 54 the evaluation never moves more than about a fifth of a pawn either way, through a queen trade, a rook trade, a pawn sacrifice and a pawn trade. Trading was the right plan. One specific trade was wrong. ✓ Verified

The other angle is the attacker's. Gukesh twice turned down trades: 28...Qb5 instead of a queen trade on d4, and 36...Rb1+ instead of 36...Rxc4, the rook trade Stockfish slightly preferred. Engine numbers do not measure how hard a position is to play. A pawn-up rook and bishop ending gives the defender a new problem every move; a pawn-up bishop ending gives almost none. Keeping pieces on was the attacker's version of the right exchange. ◐ Inferred

Part 9How to choose the right exchange in your games

• What is left after the trade, and is it good for me?
• Can my opponent force a second trade?
• If it is a pawn ending: who has the opposition, whose king is closer?
• Am I trading the opponent's best piece or my own?
• Does the trade make my opponent's job easier or harder?

For two more examples of choosing what to keep, see the bishop against knight ending Fischer steered into and how opposite-coloured bishops can still win when pieces stay on.

Your turnPractise it

Set each position up, decide your plan before you look, then play it out against Stockfish to see whether it holds.

Part 10Takeaways

• An exchange is as good as the position it leaves. Judge the result, not the trade.
• Trades come in chains. If a trade lets the opponent force another one, judge where the chain ends.
• The last piece trade is the dangerous one. Pawn endings are decided by tempi and the opposition; calculate them or avoid them.
• Defending by trading is a sound plan, as long as each trade is checked; Ding's trades from move 19 to move 54 were all fine.
• Attacking, keep the pieces that ask questions. Gukesh declined trades that made White's task easy and waited for the one that won.