--- title: "Time Pressure in Chess, Measured: 356,000 Blitz Moves" description: "Stockfish on 356,000 blitz moves by Nakamura, Carlsen, GothamChess and Botez: how big mistakes rise as the clock runs down, from 90 seconds to under 10." canonical: "https://trychessgoat.com/improve/time-pressure-data/" last_updated: 2026-10-09 --- # Time pressure, measured: what happens to top players’ moves as the clock runs down By Chess Goat · Updated October 4, 2026 Time pressure makes even the best players blunder several times as often: with under 10 seconds left, Magnus Carlsen makes 6.6 times as many big mistakes per 100 moves as with more than 90 seconds, and Hikaru Nakamura 6.4 times. We ran Stockfish on 356,372 blitz moves by four well-known players to measure it. ## How much does time pressure increase blunders? Time pressure raised the big-mistake rate of all four players we measured, by 1.9 to 6.6 times between more than 90 seconds and under 10 seconds left. A big mistake here is a move that drops the player’s winning chances by 20 percentage points or more, using the win-percentage formula lichess publishes for its [accuracy metric](https://lichess.org/page/accuracy). Hikaru Nakamura goes from 0.61 big mistakes per 100 moves with time to 3.89 under 10 seconds. Magnus Carlsen goes from 0.56 to 3.71. Each cell below is one player’s rate in the moves they played with that much time on their own clock, across 356,372 moves in total. Big mistakes per 100 moves, by clock left Clock left GM Hikaru Nakamura GM Magnus Carlsen IM Levy Rozman WFM Alexandra Botez More than 90s 0.61 0.56 1.33 2.28 45–90s 1.93 1.67 3.19 4.91 20–45s 2.45 2.03 3.92 5.30 10–20s 2.87 2.56 4.05 5.14 Under 10s 3.89 3.71 4.95 4.37 Under 10s vs 90s+ **6.4×** **6.6×** **3.7×** **1.9×** ## Do strong players handle time pressure better? Strong players handle time pressure better in absolute terms, but the clock shrinks their edge. With more than 90 seconds left, the highest rate among the four players is 4.1 times the lowest. Under 10 seconds that gap falls to 1.3 times. - **Even the best collapse under 10 seconds.** Carlsen and Nakamura make about 6.4 to 6.6 times as many big mistakes with under 10 seconds as with more than 90. - **With time, the gap between levels is huge.** With 90+ seconds left, the two world-class players make 0.56–0.61 big mistakes per 100 moves; at 2,504 blitz it is 2.28. - **The clock evens things out.** Under 10 seconds the four players are much closer together than with time on the clock: in a scramble, the gap between levels shrinks. - **Not everyone is worst at the very end.** Alexandra Botez’s rate peaks at 20–45s (5.3) and is lower under 10 seconds (4.37): her error rate is highest before the scramble, not in it. ## What does other research say about time pressure in chess? Research on time pressure in chess points the same way, with a twist. Chabris and Hearst (2003) had a chess engine check the games of 23 grandmasters and found that rapid chess “led to substantially more and larger blunders than classical chess” \[1\]. Sunde, Zegners and Strittmatter (2022) studied move times in professional chess and found that “faster decisions are associated with better performance” \[2\]. Those two results fit together: time pressure causes mistakes, but a long think is also a sign of a hard position, and hard positions are where mistakes happen. That is why this page reports rates by clock band and does not claim the clock is the only cause. Our four players are a much smaller sample than either study. ## What can these numbers not tell you? - Four players, not a population. These are their own games, measured the same way; the pattern, not the exact numbers, is the point. - Low-clock moves are not random moves: they come later in the game and in sharper positions, so the clock is not the only cause. - Engine: [Stockfish](https://stockfishchess.org/) at 40,000–50,000 nodes per position, which is strong enough to judge 20-point swings, not to settle fine evaluations. ## How was time pressure measured? The time-pressure data covers each player’s latest 2,000 blitz games, downloaded from the [chess.com Published-Data API](https://www.chess.com/news/view/published-data-api). Every position was evaluated with Stockfish at 40,000–50,000 nodes. The clock is the player’s time left after the move, read from the game record, and each move is sorted into one of 5 clock bands. Data updated 2026-10-03. Analysis by Chess Goat, independent of chess.com and of the players. Each full report is public. Player Rating Games Period Full report GM Hikaru Nakamura 3,440 blitz 2,000 2026-03-11 to 2026-09-24 [Goat Report](https://trychessgoat.com/r/hikaru-85291b45/) GM Magnus Carlsen 3,394 blitz 2,000 2025-03-23 to 2026-09-24 [Goat Report](https://trychessgoat.com/r/magnuscarlsen-b4876e03/) IM Levy Rozman (GothamChess) 2,893 blitz 1,996 2024-09-24 to 2026-09-24 [Goat Report](https://trychessgoat.com/r/gothamchess-0be88a51/) WFM Alexandra Botez 2,504 blitz 2,000 2022-11-18 to 2026-09-19 [Goat Report](https://trychessgoat.com/r/alexandrabotez-57d99692/) ## Where does your clock break? Your [Goat Card](https://trychessgoat.com/u/) measures your own clock: how often you lose on time, how you score in scrambles, and when your results fall off. For the habits that fix it, read [time management for blitz](https://trychessgoat.com/improve/time-management/) or take the [clock budget lesson](https://trychessgoat.com/learn/lesson/time-management-1/). To see your own big mistakes move by move, start with a [free game review](https://trychessgoat.com/improve/free-game-review/). Free · 5 seconds · no email ### Check your own clock, free Type your chess.com username. Your card shows your time trouble, tilt and worst hours in five seconds. Want this exact analysis on your own games, with your mistakes by clock band? The [Goat Report is $5](https://trychessgoat.com/buy), paid once. See [a blitz sample](https://trychessgoat.com/r/sample-blitz/) and [a rapid sample](https://trychessgoat.com/r/sample-rapid/) first. ## References 1. Chabris, C. F., & Hearst, E. S. (2003). Visualization, pattern recognition, and forward search: Effects of playing speed and sight of the position on grandmaster chess errors. Cognitive Science, 27(4), 637–648. [https://doi.org/10.1207/s15516709cog2704\_3](https://doi.org/10.1207/s15516709cog2704_3) 2. Sunde, U., Zegners, D., & Strittmatter, A. (2022). Speed, quality, and the optimal timing of complex decisions: Field evidence. arXiv:2201.10808. [https://arxiv.org/abs/2201.10808](https://arxiv.org/abs/2201.10808) ## Keep going - [Time managementFlagger or speed-blunderer? A clock budget for each.](https://trychessgoat.com/improve/time-management/) - [TiltA stop-loss rule, and what to do instead of pressing New Game.](https://trychessgoat.com/improve/tilt/) - [Free game reviewStockfish reviews of your own chess.com games, free and unlimited.](https://trychessgoat.com/improve/free-game-review/) - [Rating not going upIs it skill, noise or tilt? How to read a flat rating line.](https://trychessgoat.com/improve/rating-not-going-up/) Chess Goat is not affiliated with or endorsed by any coach, author, channel or site named on this page. We link to them because their free material works. Quotes link to their sources.