ELO rating

#40
"the gain of using TBs up to 7 compared to not using TBs has declined to almost nothing for recent engines." * Proof?

"experiment to join ICCF and just run SF17.1 for 3min on each move and get 100% draws" * I expect 100% losses.

@tpr said in #41:

"the gain of using TBs up to 7 compared to not using TBs has declined to almost nothing for recent engines." * Proof?
https://official-stockfish.github.io/docs/stockfish-wiki/Useful-data.html#elo-gain-using-syzygy

@tpr said in #2:

There is no limit.
If you win 99% of your games against an opponent, then you get a rating 677 higher.
If he wins 99% of his games against an opponent then you get a rating 677 higher.
Carlsen has 2839 now.
If an engine beats Carlsen 99 to 1 then it gets at 2839 + 677 = 3516.
If an improved engine beats that engine 99 to 1 then it gets 3516 + 677 = 4193.
I don't understand?

#43
"I don't understand"
An engine that beats Carlsen consistently can get to 3500 rating.
Another engine that beats that engine consistently can get to 4000 rating.

@tpr said in #44:

n engine that beats Carslen consistently can get to 3500 rating
An engine who beats a 7 year old child consistently - as in every time - can get to a 3500 rating, or a 4000 rating, or a 70,000 rating - the Elo system doesn't really give useful results for complete mismatches.

#45
You cannot farm rating from one lower rated: as you go up, he goes down and it takes too many games to farm any further.
On the contrary, you can farm rating from multiple weaker players, as did Claude Goodblood reached 2759.
https://en.wikipedia.org/wiki/Claude_Bloodgood

400 to roundabout 3400. There are Just a few Players WHO got more in tacticts or bullet.

@tpr said in #46:

#45
You cannot farm rating from one lower rated: as you go up, he goes down and it takes too many games to farm any further.

With FIDE ratings you can do so until the opponent drops below 1400 and becomes unrated, or your own rating reaches 2650 (see below).

On the contrary, you can farm rating from multiple weaker players, as did Claude Goodblood reached 2759.

That was a USCF rating under a broken rating system (might be fixed in the meantime, but I know very little about USCF).

I thought that Hikaru "Mickey Mouse" Nakamura was currently doing the same with his FIDE rating, and on 2700chess.com he has gained 11 * 0.8 = 8.8 live rating points since the September list. However they have apparently made a mistake, because the current FIDE rating regulations state (https://handbook.fide.com/chapter/B022024, § 8.3.1):

"A difference in rating of more than 400 points shall be counted for rating purposes as though it were a difference of 400 points for players rated below 2650. For players rated 2650 and above, the difference between ratings shall be used in all cases"

So I think Hikaru will not gain more than 1 point in the soon to be published October list.

@zwenna said in #48:

So I think Hikaru will not gain more than 1 point in the soon to be published October list.
I don't think he really cares. The only purpose of this adventure of his is to fulfill the condition of minimum number of rated games needed to qualify into 2026 Candidates by rating. His rating is already high enough as it is and there is a good chance none of the players who don't qualify via other paths will take over him.

The more disturbing part is that this part of the rating regulations was apparently added right now and is supposed to be effective starting 2025-10-01:

https://www.fide.com/fide-council-approves-targeted-amendment-to-rating-regulation/

but for some reason it was inserted into a document named "FIDE Rating Regulations effective from 1 March 2024" as if it had been there for a year and half already.