Carlsen calls Sindarov the favourite for the 2026 World Championship — then beats him the next day

Short answer

On 10 September, on the sidelines of the Global Chess League in Bengaluru, Magnus Carlsen said there was "no question that Sindarov is the favourite" for the 2026 World Championship against D Gukesh in Geneva. On 11 September, with White, he beat Javokhir Sindarov in 36 moves in round 26 of the same league. This piece does three things: it shows the number behind the remark — Sindarov is world number 4 on 2778, Gukesh number 32 on 2703 — measures the game move by move with Stockfish 19, and explains why a rapid win says very little about a 14-game classical match.

Event: Tech Mahindra Global Chess League 2026

How this was compiled: Remark made on 10 September 2026 on the sidelines of the Global Chess League in Bengaluru and carried by the PTI news agency; the game was played on 11 September, in round 26. Clocks read move by move from the official broadcast PGN; evaluations measured by us with Stockfish 19 at depth 24 with three lines per position, always from White's point of view. The ratings quoted are from the FIDE classical list of 1 September 2026 — the ones in the PGN are rapid ratings. The match score was computed by us from the six boards, using the league's scoring (3 points for a win with White, 4 with Black, 1 for a draw).

The essentials in a few lines

• On the FIDE classical list of September 2026 the challenger stands 75 points above the champion: Javokhir Sindarov is world number 4 on 2778, D Gukesh number 32 on 2703.
• Carlsen qualified his own line: he said "defend" is a bad place for Gukesh to start from, and that an early lead changes the picture both ways — "if Sindarov is ahead early on, it's probably very, very bad news for Gukesh".
• In the following day's game the turning point was move 16: 16...c5 took the evaluation from +1.00 to +2.71 at depth 24, where the engine wanted 16...a5.
• 23.Nd6! is the move of the game: it measures +4.49 against +2.81 for the second choice, 23.Qc4 — more than a pawn and a half between first and second.
• Sindarov spent 518 of his 1,200 seconds on two moves, 210 on 10...Bxe3 and 308 on 14...b4, and 14...b4 is the engine's third choice, behind 14...bxa4 and 14...Kh8.

The remark, and the number underneath it

The remark came on the sidelines of the Global Chess League, where Carlsen plays for the Alpine APL Pipers, and was carried by the PTI news agency. It travelled as an isolated headline, but it arrived with two qualifications that change its meaning. The first is about the word defend: "I think the word defend is really not a good place to begin, because in terms of how Gukesh has played for the last couple of years, he should not be in a position where he has to defend anything." The second is about the match itself: "if Sindarov is ahead early on, it's probably very, very bad news for Gukesh; if Gukesh is ahead early on, he automatically becomes a significant favourite."

The number under the remark is easy to check and hard to argue with. Gukesh peaked at 2794 in October 2024 and has fallen to 2703 on the September 2026 list: the reigning champion goes into a title match from outside the world's top thirty. Sindarov, who earned the right to challenge by winning the 2026 Candidates in Cyprus, sits fourth and gained another point on the latest list. Both turn 20 before the match starts, which makes it the youngest title match in history.

It is worth recording what the remark is not. Carlsen is not playing the match, and said in the same conversation that after becoming a father he decided he was "never coming back to the classical World Championship". He was speaking as the world's top-rated player and as someone who has faced both men, not as an interested party in the cycle — and the following day he had the rare chance to test his own assessment over the board.

The next day's game: 1.e4 and the Arkhangelsk

The opening ran 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.O-O b5 6.Bb3 Bb7 7.d3 Bc5 8.Nc3 O-O 9.a3 d6 10.Be3 — an Arkhangelsk in which White declines the big centre and plays for slow pressure down the f-file after the bishops come off. Sindarov took on the trade with 10...Bxe3 11.fxe3, and that decision alone cost him 210 seconds, seven per cent of his budget for the whole game.

The first moment that asked a real question came at move 14, after 13.Nh4 Ng6 14.Nf5. Black has three sensible continuations and the engine separates them by almost nothing: 14...bxa4 15.Bxa4 Rb8 is +0.34, 14...Kh8 is +0.35, and 14...b4, played in the game, is +0.55. Sindarov spent 308 seconds — five minutes and eight — choosing the third. The cost in evaluation is negligible; the cost on the clock is not.

It is the trait that shows up in all three icon-board games we have analysed in this league, his win over Firouzja in round 20 included: Sindarov concentrates time on a few decisions and plays the rest quickly. Against Firouzja that worked, because his opponent did the same and erred first. Against Carlsen the arithmetic was different — at move 16 he had 8:51 on the clock against his opponent's 12:33.

16...c5: opening the centre for the wrong side

The diagram position is instructive because the move played looks like the most natural thing in the world. Black has pawns on c6, d5 and e5 and an advance available; pushing the c-pawn gains space and frees c6 for the bishop. What the move does not address is the real source of White's edge: the knight on f5, which by now cannot be driven away, and the b3-bishop eyeing f7.

The three options Stockfish 19 considers here are 16...a5 (+1.00), 16...Kh8 (+1.27) and 16...Ne8 (+1.49). None of them equalises — the position was already uncomfortable after 15...d5 — but all of them keep the centre closed, and that is what matters: with pawns fixed on d5 and e5 the f5-knight has no target. After 16...c5 17.Qf3 c4 18.dxc4 dxe4 19.Qf2 Black is left with two loose pawns on e4 and e5 and no piece covering the king.

The broadcast's automatic annotation flags the same move with smaller numbers: 0.83 before, 2.19 after. Our measurement, at depth 24 with three lines per position, gives 1.00 and 2.71. The direction is the same; the magnitude is not. Where the reading genuinely changes is further on, at move 23.

23.Nd6!: the difference between +2.8 and +4.5

Three moves earlier Sindarov had assembled his only serious try: queen on g5, knight on g4, pawn on h5 and the idea of 24...Qh4 followed by a sacrifice on g3 or h2. The position was winning for White, but winning the hard way — with the white king airy and both clocks short. It is the kind of position where one loose move hands back half the advantage.

23.Nd6 settles it in one stroke. After 23...Rxd6 24.cxd6 the d6-pawn cannot be taken, the d8-rook has left the board, and the white queen on e2 picks up the defence of the second rank. Black's counterplay still exists — 24...Qh4 25.h3 Qxg3 26.hxg4 Nh4 is all forced, and Sindarov played those three moves in four seconds combined — but it leads nowhere: after 27.Kh1, Black's best resource, 27...Kh7, still measures +5.33.

Here the gap with the broadcast's automatic annotation is wide enough to change how the game reads. It gives 2.67 after 23.Nd6 and 2.27 after 27.Kh1; we measure +4.49 and +5.33. On our numbers the game was technically over at move 24, and what followed — including 27...Nf3, which Sindarov thought about for 134 seconds and which the engine does not even list among its top three — was the management of a lost position with under two minutes on the clock.

The clock: 518 seconds on two moves

The Global Chess League plays 20 minutes with a 2-second increment only from move 41 — in practice, no increment at all, since almost no game gets there. Every second spent comes out of the budget for every other move, and this game ended at move 36 with Sindarov on 24 seconds and Carlsen on 1:09.

The two sides of the table below show two different ways of running the same clock. Carlsen's largest single expense was 148 seconds on 22.Rae1, a preparatory move; the next two produced the two moves that decided the game — 103 seconds on 17.Qf3, straight after his opponent's error, and 82 on 23.Nd6. Sindarov front-loaded his time into decisions the engine considers near-equivalent, and reached the hard part of the game with the worse clock.

It is the same reading we gave of his win over Firouzja on day five, with the roles reversed: there it was Firouzja who spent 870 seconds on four moves, and Sindarov won the game on the move where his opponent had no time left to calculate. Anyone who wants the method behind this reading will find, in our clock guide, how to measure their own spend per move straight from the PGN.

What the game says, and does not say, about Geneva

The match runs from 25 November to 15 December 2026 in Geneva, over 14 classical games; the first player to 7.5 points wins, and 7-7 goes to rapid and blitz tiebreaks. It is worth checking the date at the source: several stories published on 10 September, all off the same agency wire, gave "24 November to 12 December"; FIDE's own page says 25 November to 15 December.

What survives from the 11 September game is a stylistic trait, not a forecast. Sindarov walks into positions where the decision is one of temperament rather than calculation — declining a trade, opening the centre, choosing between three near-identical moves — and resolves them by spending heavily and early. At rapid, that left him without a clock against Carlsen and won him the game against Firouzja. In a 14-game classical match, with a team behind the preparation and hours per move, the same trait weighs far less.

The most useful caveat is still Carlsen's own, and it points the other way: the gap between the two is wide today, but "it probably won't take a lot to potentially place some doubts in Sindarov's mind and some confidence in Gukesh". It is the one part of the remark that the rating cannot measure — and the part that decides matches.

Frequently asked questions

Update log

• 2026-09-11 — First publication, with the remark checked against the PTI wire, the game read from the broadcast PGN and the evaluations measured by us at depth 24.

Sources

• Geneva to host FIDE World Championship Match 2026 — FIDE
• Gukesh has everything to gain in title defence: Carlsen (PTI wire) — Telangana Today
• Alpine APL Pipers - FYERS American Gambits, round 26, with move-by-move clocks — Lichess
• FIDE ratings - September 2026 — ChessBase

Related guides

• An open games repertoire: meeting 1.e4 with 1...e5 without memorising Ruy Lopez theory
• How to avoid time trouble in chess: five rules for the last three minutes
• How to study chess openings with an engine without becoming a slave to the evaluation bar