Gukesh vs Praggnanandhaa, Norway Chess 2026 Round 9: full game breakdown

On the evening of June 4, 2026, R Praggnanandhaa sat down across from his long-time rival D Gukesh for the second time in a week. Three days earlier he had been fighting to stay on the board. By the time this game ended, he had beaten Firouzja, Carlsen, and Gukesh in a row and was half a point off the lead with one round to play. The win over Gukesh did not just decide a game. It turned Norway Chess 2026 into a two-horse race, and the next morning Praggnanandhaa finished the job by beating Vincent Keymer to lift the trophy outright.

If you only know the result, the headline reads "Pragg beats Gukesh again." The game itself is more interesting than that. It was a Nimzo-Indian where Gukesh walked into a deep line, Pragg found a quiet improvement at move 13, and the whole thing turned on one ambitious pawn push on move 14. The rest was technique.

This article walks through the game move by move, in plain language. No engine lines you have to set up a board for. Just the ideas, the choices, and the moments that decided it.

The setup: what was at stake in Round 9

Norway Chess 2026 is a six-player super-tournament, played as a triple round-robin with a three-points-for-a-win, one-point-for-a-loss armageddon decider after every classical draw. The time control is unusually harsh. There is no increment before move 40, and only ten seconds per move after. The format punishes slow play and rewards anyone willing to take real risks.

Going into round 9, the standings were tight at the top. Wesley So led on 14.5 points. Pragg was in striking distance at 14, with Magnus Carlsen and Gukesh trailing. Gukesh, the reigning world champion, was having a tournament he would rather forget. He had drawn four and lost four of his first eight classical games, and the armageddons were not saving him. He was on 8 points going into the game.

Pragg, by contrast, had been climbing all week. After losing two in a row in rounds 5 and 6, he had rattled off wins over Firouzja and Carlsen. Beating Gukesh in the classical game on Thursday would give him three classical wins in a row, the same Carlsen, Firouzja, Gukesh hat-trick. In the post-game interview he called Gukesh "a perfect opponent. He never backs down from a fight, no matter how the tournament is going."

The opening: Nimzo-Indian Saemisch, reaching a known line

Gukesh opened with 1.d4, which is his most common first move at the top level. Praggnanandhaa replied 1...Nf6, the Réti-style setup he has been working with for years, and 2.c4 e6 nudged the game into a Nimzo-Indian shape. 3.Nc3 Bb4 pinned the knight, and 4.a3 forced the question. White is offering the bishop pair in exchange for the doubled c-pawn and the half-open b-file.

This is the Saemisch Variation of the Nimzo-Indian, ECO code E24. The deal is straightforward. Black gives up the two bishops and gets back a permanent structural plus in the form of White's tripled-ish pawn centre on c3 and c4. For decades, strong players have argued about who is really getting the better of the exchange. The accepted view in modern practice is that Black's position is at least fine, and often easier to play for a win, especially in long games.

Gukesh continued with 5.bxc3, the standard recapture, and Pragg chose 5...b6 rather than the more popular 5...0-0. The move 5...b6 prepares Ba6, an unusual idea. The point is to develop the bishop before committing to a kingside castle. In many Saemisch lines, Black wants to play ...Ba6 only after White has committed to e4, because the bishop is most active when it is aimed at White's centre.

White played 6.f3 and 6...Nc6, hitting the d4-pawn. White responded with 7.e4, the main move, and Pragg rerouted his knight to the rim with 7...Na5. The bishop came out next: 8.Bd3 Ba6. Now White's centre is locked in place, and Black's plan is simple. Play ...c5, break in the centre, and use the a5-knight to challenge the c4-pawn.

9.Qe2 was a useful waiting move, supporting a future e5-break and keeping options open on the kingside. Pragg played 9...c5, 10.Nh3 (an unusual square, but White wants to redeploy the knight to f2 to support the e4-pawn), 10...cxd4, and 11.cxd4. After 11...Rc8 the players had reached a position that had been played in top-level games before. Gukesh, playing White, was working from a known road map. In the post-game booth, Pragg half-remembered the source: "This is the game between Nodirbek and Wei Yi from one of the Wijk aan Zees, but I cannot fully recollect." He remembered enough. He had a new idea at move 13.

Move 12 and 13: where Pragg deviates

Gukesh played 12.e5, the standard central break, and Pragg retreated the f6-knight to g8 with 12...Ng8. The retreating knight looks strange, but it is a known motif in this line. The knight is heading to a better square via e7 or h6, and the f6-square is unavailable while the d3-bishop is staring at it.

13.Nf2 was White's way of recapturing the plan from Abdusattorov versus Wei Yi at Tata Steel 2025. That game continued 13...d5 14.exd6 Nf6 15.Bf4 Bxc4 16.O-O O-O 17.Rac1 and ended in a draw. Pragg wanted more than a draw. He had a prepared improvement.

He played 13...f5.

This is the novelty. The move freezes White's e4-e5 pawn chain in place. It gives Black a long-term plan: put the knight on h6, plant it on f5 or g4, and attack White's centre from the wings. GM David Howell, commentating for Chess24, called it "an improvement over that game" and noted that Pragg had found it without checking with his second. "I don't remember anything right now," Pragg said in the post-game booth. The move was good enough on its own.

Move 14: the mistake that ended the game

Gukesh's 14.g4 is the kind of move that looks terrifying for the opponent. It attacks the f5-pawn, kicks the knight, and seems to underline White's space advantage. The problem is that it also opens the g-file against White's own king and weakens the dark squares around it. After the game, Pragg described it as "wild" and said, "I don't believe in this too much for White, but it is complicated."

Praggnanandhaa was being polite. The move is not complicated for Black. It is losing.

Pragg's response was 14...Nb3, grabbing the a1-rook. The tactical point is that 15.Rb1 trying to save the rook runs into 15...Nxd4, and after 16.Qe3 Rxc4, White is losing material without compensation. Howell showed a clean line where 17.O-O (forced, since 17.Bxc4 Nc2+ wins the queen) 17...f4 18.Qe4 (or 18.Qxf4 Ne2+) 18...d5 wins the queen by force. White's attack on the kingside is too slow.

Gukesh instead tried 15.Bb2, accepting the loss of the exchange, and after 15...Nxa1 16.gxf5 he had given up a whole rook for a pawn and the beginnings of a kingside attack. Black played 16...Nb3, retreating the knight to safety, and 17.Rg1 Nh6. From here on, Pragg was a rook up for almost nothing, and the rest of the game was a question of whether he would let the clock get in the way.

He did not let the clock get in the way. In his first game against Gukesh the previous week, Pragg had been winning and then collapsed in a time scramble. He said later, "I thought I managed my time well and I was also calculating well. I am quite proud about this game, honestly, because this is how I usually lose to him."

Moves 18 to 27: Gukesh fights, Pragg finds only moves

Gukesh was not going quietly. 18.f6 was his only practical try. 18...gxf6 19.Ne4 fxe5 20.Nd6+ Kf8 21.Qxe5 Qh4+ 22.Kd1. The white queen is suddenly threatening Qg7 mate, and the black king looks exposed. Pragg found the only correct defensive idea: 22...Rg8, defending g7 while keeping the king safe on f8. After 23.Rxg8+ Nxg8 24.Nxc8 Bxc8 25.Kc2 Na5, the position simplified into a technically won minor-piece endgame for Black. White is a piece down but has two pawns and active pieces.

Gukesh kept trying. 26.d5 opened the centre and threatened to bring the queen back into the attack with Qg7+. Pragg played 26...Qh6, defending the threat, and 27.Kb1. Gukesh, in a lost position, kept the game going. The commentators noted that, objectively, Gukesh "even got himself back into the game at one point." Pragg never let him back in.

Moves 28 to 34: Pragg converts

The technical phase was clean. 27...exd5 28.cxd5 d6 29.Qg3 Nb7 30.Bc1 Qg7 31.Qh4 Qf7 32.Ba6. White's last try was 32...Qf5+ 33.Ka2 Qxd5+ 34.Bc4 Qf5, and the game was over. Praggnanandhaa had a piece and a king position. White resigned.

The clock showed 34 moves. The full game had taken less than two and a half hours of clock time. Gukesh fought, but the position never gave him a real chance.

Why this game matters beyond the result

Three things make this game worth studying, even if you never play the Nimzo-Indian.

The first is the opening. 13...f5 is the kind of move a club player can borrow directly. In any Nimzo-Indian where Black has played ...c5 and ...cxd4 and White has e4 and e5, the ...f5 break is a useful way to slow White down. It is not a one-game novelty. It is a plan, and it works at any level.

The second is the psychology. Pragg later admitted that this was the kind of position where he normally overcomplicates and loses. He deliberately played simple, fast, technical moves. That is a transferable habit at any level. When you are winning, do not try to be clever. Trade pieces, activate your king, and convert.

The third is Gukesh's mistake. 14.g4 is the kind of move that wins at club level and loses at super-GM level. The difference is that strong players calculate one move further. At move 14, Gukesh was thinking about the kingside attack he was starting. Pragg was thinking about the queen on a8. That is the entire gap between two young Indian super-grandmasters, and the gap is not as wide as the Elo difference suggests.

What you can take to your next game

If you play the Saemisch as White, the lesson from this game is straightforward: do not play g4 just because you can. The move invites a knight to a permanent home on h6, where it will live for the rest of the game. White should consolidate first, complete development, and only then think about kingside pawn pushes.

If you play the Nimzo-Indian as Black, the position is not as drawish as the reputation suggests. Black has real winning chances in the Saemisch, especially when White plays for f3 and e4 without thinking about the long-term consequences.

Either way, Pragg's hat-trick in the last three rounds is one of the great comeback runs in recent super-tournament history. He was effectively out of it after round 6, with two classical losses in a row. Five days later he had won the tournament. That is the kind of run that does not happen by accident.

The wider picture

Gukesh finished Norway Chess 2026 in last place, on 8 points. He lost five classical games, won one, and drew four. He dropped 14.8 rating points and fell to 26th in the world on the live list. For the reigning world champion, it was a difficult event to digest. He will almost certainly come back stronger. He usually does.

For Praggnanandhaa, this was a statement. He had been written off in some quarters after a tough Candidates. Norway Chess 2026 was the answer. Win three games in a row against Firouzja, Carlsen, and Gukesh, all with classical time controls, and the conversation changes. The gap between the top two Indian players is, on this evidence, very small.

Norway Chess 2026 is over. The next big event is the next Candidates cycle. If both players make it to the same Candidates, the chess world will be watching. For now, the only sensible prediction is that the next time these two sit down across from each other, neither of them will be overconfident.