The group affected consists of the world’s elite grandmasters, exactly 69 of whom are currently to be found in the 2650+ bracket (Alexander Morozevich is exactly 2650 and so presumably does not count). The rule aims at preventing very strong grandmasters from gaining points excessively when playing against much lower-rated opposition. Beating amateurs should not be evaluated as though they are just 400 points weaker. The rating calculations will be based on the true, potentially much larger, difference.
FIDE CEO GM Emil Sutovsky explained the intention of the change on X (formally Twitter):
No more farming. If you are a 2650+ player, do prove your skill vs opponents of comparable strength.
Why 2650? It is a top-100 level, and these players rarely face low-rated opponents. While, for example, GMs with 2500-2650 mostly play in large opens, and they should not suffer from the new regulations.
Additionally, FIDE prepared a detailed questionnaire related to ratings. After what period a player should be considered inactive? How to approach rating decay, etc.
This will be sent to the top – 100 and women’s top-50, to have their feedback before taking the decisions on a whole range of matters.
Chess.com tells us that “the timing of the change has raised eyebrows, as it comes just weeks after Nakamura gained nine rating points from his 11 wins against players rated as low as 1800 in the Iowa Open and the Louisana State Championship this month.”
Emil Sutovsky tweeted:
Just to be clear.
The rule is not about Hikaru. He did trigger it, but when we started to dig, it turned out there were at least five players more of 2650+ level who, in 2024-25 regularly played events with a string of very low-rated opponents, abusing 400-points rule.
It was…
— Emilchess (@EmilSutovsky) September 29, 2025
Sonas on the rating change
Leading chess statistician Jeff Sonas, who did a lot of analysis for FIDE on rating problems (and has written numerous articles for us), helped understand the current situation. This is what he told us:
The 400-point-rule ensures that a player’s Elo expected score (based on rating differences) can never be lower than 8% or higher than 92%, even for extreme Elo differences. Thus a victory by the higher-rated player in a game would always bring at least +0.8 Elo points (that’s 10 times 0.08, for a GM with a K-factor of 10), under the 400-point rule.
With the removal of the 400-point-rule for the highest-rated players, their Elo expected score can now be as high as 99% or even 100% against the lowest-rated opponents, making it much less likely for the GM to be able to gain any significant rating points from a series of games from Open events.
Say you have a 2650 top GM facing a 1900-rated player. The rating difference of 750 points means that the expected score for the GM would be exactly 100%. Thus a win would leave the GM’s rating unchanged, a draw would lose them 5.00 Elo points, and a loss would lose them 10.00 Elo points (check out the FIDE Handbook here). This is a clear disincentive for elite players to take part in Open events.
Could we have predicted what would happen when Hikaru Nakamura played in an Open event? Admittedly there is very little data on what’s happened in the past, when a 2800+ elite GM faces much weaker opponents who are rated approximately 800 Elo points lower; it simply hasn’t happened very much.
But we can come close. In some research I did a couple of years ago regarding the 400-point-rule, I looked at the actual numbers from when players rated 2600-2799 Elo faced opponents rated below 2000 Elo. During a 15-month period from January 2022 until April 2023, there were 199 such games, and the higher-rated players scored 192/199 (96.5%) in those games.
So every month, there were approximately 13 games being played by 2600-2799 Elo players against sub-2000 opposition, with an average score of 12.5/13. Let’s pretend that each month, there was one player who played in a 6-round Open and a 7-round Open. What would we expect to happen?
We know that on average (from real-life numbers), the GM would score 96%-97% in such games, and so a typical total score for the GM (across the two events) would be 12.5/13.
With the 400-point rule in place, the Elo expectation would be capped at 92%, and so their total Elo expectation across those 13 games would be 12/13. Overall the GM would have gained 5 Elo points total, if they actually scored 12.5/13 (and the 400-point-rule was still in place).
Now that the 400-point rule will not be used for players rated 2650+, their Elo expectation in this scenario would be 13/13. And so instead they would lose 5 Elo points if they managed a typical performance of 12.5/13.
My biggest concern about the new rule is that it introduces a discontinuity around 2650 Elo. So if you are rated 2650 and you play 13 games in those two Opens, you can expect to score 12.5/13 and gain 5 Elo points. But if you are rated 2651 and you play 13 games in those two Opens, you can expect to score 12.5/13 and lose 5 Elo points. Nevertheless, I don’t have a big problem with what FIDE did with this latest change.
Possibly it would have been better to say that in this scenario you get a 500-point-rule or a 600-point-rule instead, so there’s something for the GM to play for (you could still gain rating points with a perfect score). And I think in that case, the Elo expected score of 96% (for a 500-point-rule) or an Elo expected score of 98% (for a 600-point-rule) probably would match reality a little better than what will happen now (essentially an Elo expected score of 100%).
Jeff Sonas is a statistical chess analyst who invented the Chessmetrics system for rating chess players, which is intended as an improvement on the Elo rating system. He is the founder and proprietor of the Chessmetrics.com website, which gives Sonas’ calculations of the ratings of current players and historical ratings going back as far as January 1843. Sonas graduated with honors with a B.S. in Mathematical and Computational Sciences from Stanford University in 1991.
