Fide rating system changes

You know how old timers always say that players seem much stronger than their ratings would suggest, compared to back in the day? Ever been beaten by a scholastic player who seemed to be dramatically underrated?

Well, good news everyone who isn't aware! Fide has proposed changes to their rating system to improve rating accuracy and combat rating deflation. The decisions will be released in December, and go into effect in January.

https://fide.com/news/2538

The main proposals are to raise the minimum entry rating from 1,000 to 1400, and raise all ratings proportionally for current players under 2000 (85% of all players) by 0-400 points.

For example, 1000 Elo players will become 1400s, but 1900s will become 1940s. The formula is given in the pdf. Since 1000 strength players will become 1400s, this change is only immediately numerical in nature, you won't need to become stronger in order to get a Fide rating.

This is intended as a one-time measure to (fairly) compress players' ratings together, so that the game results between players of different skill levels will become a more reliable indicator of their respective strengths. This should bring statistics into line with the originally intended spread of the Elo system- for example, instead of a 1500 beating a 2100 15% of the time, this should come down to 2%.

This is all intended to reverse and then help prevent deflation caused by new, low-rated players who are improving rapidly, reducing others' ratings unfairly. Many of these players end up quitting chess in their teen years, draining a huge number of points from the system permanently, and even if they stick around, it's deflationary. The original underrated new players win a large number of points from opponents who then themselves become underrated, and the deflation spreads out from there.

Adjustments to the formula used to calculate initial ratings are also proposed for improved accuracy (currently a vast number of beginners are given far too low an initial rating), and this combined with K factors remaining at 40 for under 18s and all players with less than 30 rated games, should also contribute to tackling the issues caused by those improving faster than their own rating can keep up.

Fascinatingly, since titled players are generally the furthest removed from the underrated newbies in this sequence of who-plays-who, they're basically unaffected by the current deflation, and at the top levels there is actually too much rating compression. To that end, the '400 point rule' is proposed to be changed back to its original format.

Currently when a player beats somebody rated more than 400 points below them, they're allowed to ask, only once per tournament (for their game with the greatest rating disparity), that the rating changes be treated as if the difference is capped at 400. For example, a 2400 rated player would usually win only 0.1 rating points from a 1400. But this would allow the rating change to be treated as if the 1400 is a 2000, resulting in a 1.8 point win.

Under the proposals, the rule will be reverted so that this applies for ALL games with more than a 400 point rating difference. This would make tournaments with unequal pairings more worth their time, and whilst this would technically allow for 'farming' of low rated players, ain't NOBODY GOT TIME FOR THAT. It's not realistic to think that high rated players would bother. But it should raise the rating ceiling a little, allowing for decompression at the higher levels, and if it encourages strong players to play against significantly weaker players on occasion, the rating loss is still small and not a huge blow for the lower rated, and they get the learning experience.

Additionally, if the 1400 beats the 2400, it would still be treated as if the rating difference is 400- but this only changes the rating win from +19.9 to +18.2, so an upset is still an upset and this isn't going to penalise the lower rated players who have a game with outstanding performance.

As a side note, for chess.com users, this means that your online rating, which is currently slightly inflated for those under 2000, should become more closely comparable with your likely Fide rating. Ain't that handy?

TL;DR:

They're compressing ratings under 2000 together, and making 1400 the new entry point. Ratings will be much more accurate from the very start and kids + rapid improvers won't suck away everyone else's points. Strong players will win more points from games against opponents over 400 points lower rated than them. Your Fide will match your chess.com rating better.

Will be interested to see if this works.

Me too. I deep-dived into the analysis pdf given by their experts, it looks like it holds up but we probably won't know for sure for years.

I hope getting FIDE rating and increasing it becomes easier because where I stay it's hella hard to do that.

It should help in terms of not losing hella points to underrated people, which means you don't have to spend forever winning your points back just to get where you were. And since your opponents won't be losing all their points to underrated players and becoming underrated themselves before they play you, if you lose to them you'll lose fewer points than you might've. Basically, you can expect less fluctuations. So if you're improving, you should have an easier time increasing your rating, yep.

Sporkled do you play otb?

Though, if it's hard to get a rating somewhere, the decrease in rated players will make it even harder in those areas.

If the minimum entry rating requirement becomes 1400, I will never have a chance to become a rated player. I'm below 1000 as it is right now, so once the rating increase of the 1000 players to 1400 happens, it would be completely out of reach for me.

No because "Since 1000 strength players will become 1400s, this change is only immediately numerical in nature, you won't need to become stronger in order to get a Fide rating."

Currently, 1000 Fide = around 1400 chess.com. Today, you'd have to be around 1400 strength here before you'd qualify for a rating.

Fide will bring existing members with a Fide rating of 1000 up to 1400, and make that number the new bar for entry. The actual strength of entry level players doesn't change, just their number, and when you play games to get a Fide rating, you will need to be that strength.

So instead of a chess.com 1400 being comparable to a 1000 Fide, they will be comparable to a 1400 Fide, because the numbers changed. Whether today, next week, or after this change happens, if you wanted to get a Fide rating you'd need to be around 1400 on this site. You wouldn't be able to qualify for a rating at 1000 chess.com strength.

Actually nope. I'm starting to consider it, which is how this caught my eye. My local chess club has a quite famous author GM, it's very tempting to start attending.

WOW amazing!!!!
We have an IM (Carissa Yip ) but no GM as of rn

Rating above 2000 tempts to be a little bigger and rating below 1800 tempts to be smaller because players are amateurs still

There may be some farming due to the 400-point max difference for calculation purposes. That existed last century for USCF when 350 was the max calculated difference (2 points for a win if you were under 2400 and 1 if you were over). You would occasionally see people that were close to their rating target enter a few small tournament where they were the highest rated and play only one game (garnering 2 points) before withdrawing. If they were at 1994 and wanted to be 2000 so they could charge more for lessons then three tournaments was enough to get there.

I will grant that such players are also more likely to stop rated OTB play once they reach that magic number (can't risk dropping under 2000 when charging for lessons) so the farming effect may be more limited than even the limited number of players would suggest.

To combat natural deflation USCF does ratings tournament by tournament (not game by game) and provides bonus points for players having much better tournaments than their rating would suggest (such as doing something like doubling the gains for every point over 40 that was gained in a five round tournament - I'm not sure what the current bonus point tripwires are). That increases the speed that improving players have their ratings catch up to their strengths. Also USCF has rating floors that a player cannot drop below (originally emplaced to combat rating manipulation to sandbag and win class prizes - still its primary function) and for older but still active players that may introduce a tiny bit of counteracting inflation to the system as their floor exceeds their current strength.

The USCF rating committee monitors the average rating of 35-45 year old OTB players (players age into and out of that range but the average rating for the range is what is monitored) and keeps that consistent by adjusting things like the tripwire number for bonus points. When the average rises too high it raises the tripwire to reduce bonus points, and when the average drops too low it reduces the tripwire to increase bonus points.

That said, a 2000 rating today is automatically stronger than a 2000 rating some years back simply because there has been more research and learning done that players have access to. A solid 2000 rated player from 1970 transported in time to 2023 (and given no opportunity to study the advances in theory) would quickly lose rating points until learning enough to stop the hemorrhaging.

Nice article. A couple of points of (hopefully useful) feedback and a question if I may. I haven't read all the comments so you may have received similar already.

1. You should put the TL;DR at the start. I read your full article before I saw it, which was fine in this case but I think readers would ideally want to see it first.
2. I don't agree that Chess.com ratings will more closely reflect FIDE ratings as a result of these changes because, amongst other things:
   1. Chess OTB is quite different to online, particularly for those who don't play OTB regularly, e.g. 3D vs 2D, a real opponent opposite you, no assistance from arrows/valid moves/analysis board for (lazy/tired) calculation.
   2. The pool of players online is very different to those who will play FIDE-rated events in your area, e.g. in the UK FIDE-rated events are still fairly hard to access unless you live in/around London, despite recent efforts to improve things.

My question relates to my son's FIDE rating... he played a (very rare) FIDE rated junior blitz tournament at the weekend and he now has 10 FIDE-rated blitz games, 2 of which are from 2021 (the last time he was able to play such an event!). I want to calculate his initial rating (which is pretty simple in itself) but I'm not sure whether the ratings of his two 2021 opponents should be adjusted with the new formula or not before calculating his rating. Do you happen to know? FIDE's online tool just requires you to enter the average rating of rated opponents (Rc), so that doesn't help.

Kind regards,

Jules