Twenty minutes with no increment: what 6,199 Global Chess League moves say about the clock
The league's time control only gives an increment from move 41. We measured the rate of serious mistakes by time band across the first ten rounds, and the curve is steeper than intuition suggests.
10 min read · Published on 2026-09-07
Short answer
The Global Chess League 2026 is played with 20 minutes and a 2-second increment only from move 41. We took the 6,199 clock-stamped moves of the 60 games of the first ten rounds and measured how often a serious mistake occurs by remaining time: 0.4% above five minutes, 2.2% between two and five, 3.0% between one and two, 3.5% between thirty seconds and a minute, and 4.1% below thirty seconds. This article explains what the curve means in practice and what to do with it in your own time control.
Event: Tech Mahindra Global Chess League 2026
How this was compiled: Survey of the 60 games of the first ten rounds of the Global Chess League 2026, covering 6,199 moves with clock times.
The essentials in a few lines
- With more than five minutes on the clock, only 0.4% of the moves in the 60 games were classified as serious mistakes; below thirty seconds, 4.1% — ten times as many.
- The decline starts before visible time trouble: between two and five minutes left, the rate is already five times that of a comfortable clock.
- Forty-seven of the 60 games reached move 40, and the median time left at that move was 80 seconds — the increment arrives when there is barely any time left to use it.
- Mistakes with time on the clock also happen: Anand handed Sindarov the game with four minutes available, against his opponent's two minutes and forty.
- The classification of serious mistakes is Lichess's automatic analysis of the official broadcast; the curve shows a tendency, not a verdict on individual moves.
The time control, and why it is not an ordinary 20+2
How does the Global Chess League 2026 time control work?
Each player gets 20 minutes for the whole game and receives 2 seconds per move only from move 41. There is no increment at all in the first forty moves: the clock only goes down.
The difference from a conventional 20+2 is larger than it looks. With an increment from move one, a player who averages ten seconds a move recovers a good part of what was spent and reaches move 40 with a cushion. Without it, every second spent thinking is a second lost, and the opening competes directly with the endgame for the same budget.
The design is deliberate: it forces quick decisions in the phase where most players would like to think, and returns a survival minimum at the moment positions tend to turn technical. The cost is that players arrive at the technical phase with the clock spent — and that is where the league produces its swings.
For online players, the closest equivalent in feel is 10+2 or 15+0, not 20+2. The number that matters is not the starting total but how much time is left per move in the phase where the game is decided.
The curve of the mistake
How fast does move quality fall as the clock tightens?
It falls before time trouble is visible. Across the 6,199 measured moves, the rate of serious mistakes is 0.4% above five minutes, rises to 2.2% between two and five minutes and reaches 4.1% below thirty seconds — ten times the comfortable rate.
The most useful part of the table is not the extreme but the middle. Common intuition treats time trouble as a phenomenon of the last thirty seconds, but the big jump happens between the band above five minutes and the two-to-five band: from 0.4% to 2.2%, a more than fivefold rise with time most players would still consider comfortable.
From thirty seconds down the curve keeps rising, but more slowly — 3.0%, 3.5% and 4.1% in the last three bands. That fits what players observe: with very little time, calculation stops and moves come from patterns, and a pattern errs less often than a calculation cut off halfway.
One methodological caveat the data itself imposes: the classification of serious mistakes is Lichess's automatic analysis of the official broadcast, applied to rapid games. It serves to compare bands with each other — which is what this survey does — not to judge an individual move. Where this article states the value of a position, that evaluation was measured separately with Stockfish 19 at depth 24.
| Time left after the move | Moves | Serious mistakes | Frequency |
|---|---|---|---|
| Above 5 minutes | 3,387 | 14 | 0.4% |
| Between 2 and 5 minutes | 918 | 20 | 2.2% |
| Between 1 and 2 minutes | 470 | 14 | 3.0% |
| Between 30 seconds and 1 minute | 313 | 11 | 3.5% |
| Below 30 seconds | 1,111 | 45 | 4.1% |
Move 40 arrives with 80 seconds
Does the increment from move 41 solve the problem?
Not for anyone who has already spent the budget. Of the 60 games, 47 reached move 40, and the median time left at that move was 80 seconds. Almost half the players got there with under a minute.
Eighty seconds is what separates most players from the move where the increment begins. Two seconds a move keeps someone alive indefinitely, but they do not pay for calculating a rook endgame: the player enters the most technical phase of the game with a decision budget that covers only pattern moves.
That is what the two most discussed games of the league show so far. Magnus Carlsen defended 52 of his 89 moves against Ian Nepomniachtchi with under fifteen seconds — and found the only drawing move with five seconds left. On the other side of the statistic, several games in the same event were decided by a single move played in that band.
The practical conclusion for anyone playing similar time controls is the familiar one, now with a number attached: the time budget has to be decided before the opening, not during it. Reaching move 40 with two minutes costs little in simple positions and costs the game in any endgame that requires counting tempi.
A concrete case: 36 seconds and a piece for free
What does a clock-driven mistake look like in a real position?
Usually not as an absurd move, but as a plausible one that overlooks the opponent's resource. Against Daniil Dubov, Aravindh Chithambaram played 37.Be5 with 36 seconds in a position evaluated at +0.39, and the reply 37...Rxe5 decided the game.
Before 37.Be5 the position was balanced: the engine gives +0.39 for 37.Qc1, the best move. The move played loses nothing immediately — it loses to 37...Rxe5, giving up the exchange in order to win it back with 38.Rxe5 Ng3+, a forking check that recovers the rook and leaves Black a bishop and a pawn ahead.
The detail that makes the example useful is the other side. Had Black captured with the bishop — 37...Bxe5, the natural move — the evaluation would return to zero and the game would stay level. Dubov had one minute and twenty-six seconds on his clock, three times his opponent's time, and used that difference to find the right order of captures.
That is how the clock decides games in practice: not because someone flags, but because the difference between two plausible moves requires a three-move check that nobody performs with thirty seconds. Punishing the mistake, on the other side, also takes time — a player in a hurry tends to accept the simple version of the capture.
| ♛ | ♜ | ♚ | |||||
| ♟ | |||||||
| ♟ | ♟ | ♟ | |||||
| ♜ | ♞ | ♟ | |||||
| ♙ | ♝ | ♖ | |||||
| ♙ | ♘ | ♙ | ♙ | ||||
| ♙ | ♕ | ♙ | ♗ | ||||
| ♖ | ♔ |
Full line: 1.Be5 Rxe5 2.Rxe5 Ng3+ 3.Kh2 Bxe5 4.f4 Bc3
What to do with this in your own time control
How do you manage the clock in rapid games without an increment?
By splitting the budget by phase before the game starts and checking the clock at fixed points, not once it has already tightened. In twenty minutes with no increment, a reasonable budget is five minutes to move 15, eight more to move 30 and the rest for the endgame.
The hard part is not setting the budget but respecting it, and the only way to respect it is to agree checkpoints with yourself. Looking at the clock at moves 15 and 30 and comparing with the plan is a habit that settles in within a few games; looking at the clock when it is already red is not a habit, it is a reaction.
The second decision is where to spend what is left. The survey suggests a simple criterion: spend where the position changes nature — a queen trade, a passed pawn appearing, a file opening — and save on developing moves and forced recaptures. Those are exactly the moments where the two annotated games in this coverage were decided.
And it is worth correcting the lazy explanation that follows every rapid loss. Mistakes with time on the clock exist and are common: against Sindarov, Anand handed over the game with four minutes available while the winner had two minutes and forty. Separating the two causes — lack of time and lack of judgement — is what makes reviewing a game useful.
- Set the budget by phase before the game and check it at moves 15 and 30.
- Spend time on the moves where the position changes nature, not on the forced ones.
- In endgames, count king tempi before choosing a square — that is the calculation time trouble removes.
- When reviewing, note how much time you had at the move you erred: without it you cannot know what to train.
Frequently asked questions
How do you manage time in 20-minute games with no increment?
By setting a budget per phase before the game — for example five minutes to move 15, eight more to move 30 and the rest for the endgame — and checking the clock at fixed points. In the Global Chess League 2026 data, the frequency of serious mistakes rises from 0.4% to 2.2% once the clock falls below five minutes, well before visible time trouble.
Does the 2-second increment from move 41 help?
It helps you not to lose on time, not to play well. Of the 60 games measured, 47 reached move 40 with a median of 80 seconds left — almost half the players arrived with under a minute. Two seconds a move sustain the game but do not pay for calculating an endgame.
Is every rapid mistake the clock's fault?
No. In the same event, Viswanathan Anand handed a level game to Javokhir Sindarov with four minutes on the clock, against his opponent's two minutes and forty. Judgement errors happen with time available, and telling the two cases apart is what lets you choose the right training.
How were these figures compiled?
From the PGN files of the official broadcast of the first ten rounds, covering 60 games and 6,199 clock-stamped moves. The classification of serious mistakes is Lichess's automatic analysis of that broadcast; the evaluations of specific positions cited in this coverage were measured separately with Stockfish 19 at depth 24.
Update log
- 2026-09-07 — First publication, with the survey of the first ten rounds.
Sources
Related guides
See which time band your mistakes fall in
The tournament curve applies to professionals. Yours is different, and it is already recorded in the games you have played: import them and compare mistakes by phase and by time remaining.
Create a free account