Rated Chess960 games: time for a change

@Former_Player said in #10:
> Well, I agree that online ratings are a dubious metric in general, but I've got a couple of points to consider:
> 1. Chess960 is rather a separate full-fledged kind of chess, not its variant. Moreover, if anything, the start position #518 (aka the standard one) is a variant of Chess960 ;)
> 2. In the absense of FIDE or other over-the-board ratings for Chess960, its fans have to be consent with online ratings. But what do they get? The otb ratings don't exist and the online ones are a mess. Quite sad.

These are the strongest points you have made because they show respect for 960, rather than seeming too worried about the ratings. I would personally be happy to see the ratings being broken up by time control and even given quick pairing buttons, especially in light of the seeming failure of the hourly 960 team arenas. I just have low expectations

While I understand that Lichess could hardly create too many new leaderboards, dividing the Chess960 rating lists into 3 categories (bullet, blitz and slower time controls, or something like that) would also make sense to me. A more subtle division could make it difficult to stay in the lists. In fact, staying in the (standard) rapid list is difficult and time-consuming, as one needs to play a lot, especially when the opponents are significantly lower rated. (When I set a challenge for a standard rapid game with a time control 10+0 or 8+0 within a 400 points rating difference, I often have to wait for several minutes before getting paired.) I guess that in Chess960 it would be even more difficult to stay active in a "rapid list", but at least player's ratings would reflect their strength in a given time control. It indeed looks odd to treat Chess960 hyperbullet, blitz and rapid games in the same way.
Another problem with the local ratings is that many titled players play mostly/only the titled arenas, where the ratings are underrated by 150-200 points compared to the rest of the field. It is no wonder when strong GMs and IMs occasionally enter those arenas with a provisional rating 1500 and mostly fail to get too high because the opposition is very strong.
That said, I am glad that Lichess offers comparatively many opportunities to Chess960 players. (Staging serious tournaments without berserk would be great, but Lichess does more than most of the others.)
In general, there is a striking contrast between a high number of titled players who would be glad to play Chess960, and a low number of organizers and sponsors of Chess960 events. (Now I am referring primarily to the OTB events.)

Gh6

@Former_Player said in #1:
> Hello everyone,
>
> The rating system for Chess960 here on Lichess is highly doubtful, because all chess960 games are rated in exactly the same way irrespective of the time control employed.
>
> Chess960 is (unlike all other variants present here) very similar to the standard chess where no one questions the existence of separate rating categories for separate time control ranges - 'ultrabullet', 'bullet', 'blitz', 'rapid', 'classical'. Obviously, those rating lists differ significantly, because different skills are essential for each category.
>
> Meanwhile, a 30-second Chess960 game affects the ratings identically to let's say a 30-minute Chess960 game. Does it make any sense?
>
> As a result, the ones whose mice are faster and premove skills better (and in general those who enjoy ultra-fast time controls) are regarded as the 'best' Chess960 players, whereas strong or even world top GMs who just happen to prefer blitz over bullet or ultra-bullet (and who would most likely beat the former ones in blitz matches) are struggling to earn 2500 and enter the Top 10.
>
> Thoughts?

What rating system does Lichess use? Ratings are calculated using the Glicko-2 rating method developed by Mark Glickman. This is a very popular rating method, and is used by a significant number of chess organisations (FIDE) . So please don't blame Lichess team/staff about rating .

Totally agree that the 30 seconds mouse freaks control the leaderboards with variants. When you look at their games it is often disgusting chess. In 960but also in other variants.

btw @RealDavidNavara
You do understand that in rapid games your rating is ridiculous high.
Out of the half million players in rapid this week only a bit more than 200 fit your criteria of 400 rating points difference.

already Lichess-Classic is sort of unreliable rating pool due small amount of games.
lichess.org/stat/rating/distribution/chess960
From here you can see that only 10 000 players played over a week. If divide them in subgroups iabout 2/3 those would be blitz so rapid and classic would divide rest of players over rating scale from total beginner to grandmasters. rating distribution would look even more odd than current 960 distribution.

#9:
> I agree with Former Player point of view about ratings, it's time for 960 be considered as what it is, not a chess variant, but simply chess
#10:
> 1. Chess960 is rather a separate full-fledged kind of chess, not its variant. Moreover, if anything, the start position #518 (aka the standard one) is a variant of Chess960 ;)

Because of that, my soulution would be to remove the 960 rating completely and use your standard ratings in 960 games instead.

@dampooo said in #16:
> btw @RealDavidNavara
> You do understand that in rapid games your rating is ridiculous high.
> Out of the half million players in rapid this week only a bit more than 200 fit your criteria of 400 rating points difference.
My description was inaccurate, as I either click on the 10+0 button at wait for several minutes, or set up a challenge for the 8+0 game with a rating difference up to 450 or 500 points.
That said, my local rapid rating is not RIDICULOUS(LY) high, it is appropriate given my OTB results and the fact that I play very few tournaments here (and no prize events) and berserk little. Other strong GMs like @Zhigalko_Sergei, @Durarbayli, @Rakhmanov_Aleksandr or perhaps @DrawDenied_Twitch have comparable rapid ratings here.
Next, there are many more players with high rapid ratings, just some of them are not active enough to appear in the lists. (I have already played two GMs rated here around 2550 with questions marks.) By the way, I am not on the list for the same reason, as my rating deviation has surpassed 75. It is no wonder, as after one game with a 450 points rating difference it usually drops by less that 1, while during a week of "inactivity" it usually rises by about two points.

Let's go back to the original topic. If the Chess960 ratings would be split into three categories, I would be active in bullet and blitz and less active in rapid, staying outside of the local leaderboards. That said, even then it could be good to have a more or less adequate rapid Chess960 rating.

@ShiningDrongo said in #20:
> #9:
>
> #10:
>
>
> Because of that, my solution would be to remove the 960 rating completely and use your standard ratings in 960 games instead.
And what about removing the online blitz ratings completely and playing unrated games only, when there are FIDE blitz ratings? :-D