lichess.org

Lichess should improve its rating statistics and computer analysis system!

Since we only get 25 computer analyses per day, I've been playing over at chess.com lately so that I can catch up with my Stockfish computer analyses over here, and I must say, while Lichess.org is a far superior chess experience compared to chess.com, they do actually have some perks over there that's better. For example, when running a computer analysis over there, it shows pretty much the same stuff we get here (and the black and white strength graph here is much better done, no doubt), like inaccuracies, mistakes and blunders, but their Komodo engine does show additional information, like "Brilliant", "Best Move", "Excellent" "Good", "Book Move" and "Missed Win". It'd be useful for those of us who really like to analyze our matches, if we could get additional information on how good our moves were in any given game. I'm not sure if Stockfish analyzes the quality of the moves like that though, maybe that's a Komodo exclusive, but if possible, that would be a very nice upgrade in future Lichess analyzed games.

The second request I'd like to see implemented here, is more details in the ratings statistics. Over at chess.com, they actually show how many global players there are in blitz, rapid and so on (like your position being #500,000 out of 3 million players), and what the average rating is, and more detailed leaderboards. I know we have leaderboards here too, but theirs is a little better. For example, they even show ratings statistics for correspondence, which isn't shown here, and last I asked about that, the reason given was that cheating is more common in correspondence due to its slow pace, but if that's the rationale then I don't see why classical matches should have ratings statistics, because it's just as easy to cheat in classical, given that classical matches allow enough time for cheaters to do their thing (I'm not a cheater and never will cheat, for the record, so don't get any ideas folks!).

Also, I understand why computer analyses are limited given that they're server intensive, but I suggest that for patrons, we should be allowed more computer analyses. This should really be the only extra for patrons. I wouldn't bother with chess.com otherwise, but I want all my matches analyzed, and so I have to limit my playing here in order to catch up with all these analyses.

Anyway, I really like Lichess.org and it'd be great if @thibault could improve the few missing features lacking here that's available on chess.com, keep up the good work man!

I do agree "book move" would be useful. The rest doesn't make much sense. A brilliant move (!!) is just a move that doesn't change the eval. There is no such thing as brilliant moves to engines.

Everything will be the same for everyone. That's one of the principles behind lichess. Everything is free forever and for everyone. Patrons won't get additional benefits. The amount of request analysis was quite recently changed from 20->25, maybe it's time to raise it again.

@DannyGreene123

Because if the number of analyses were indefinite it might overload the server. Or maybe just Stockfish. But either way it wouldn’t be good.

Nobody (myself included) has found a way to speed up analyses without reducing their depth/quality. Perhaps if some breakthrough is made (maybe some form of multiple-iteration analysis which analyzes more like a human does) that might reduce server load.

@EliasAlucard
> The second request I'd like to see implemented here, is more details in the ratings statistics. Over at chess.com, they actually show how many global players there are in blitz, rapid and so on (like your position being #500,000 out of 3 million players), and what the average rating is

From what I can tell it’s exactly what the rating distribution page does ;)
lichess.org/stat/rating/distribution/blitz

"Brilliant", "Best Move", "Excellent" "Good", "Book Move" and "Missed Win"
Those are all obsolete notions from the previous century. There objectively is only "mistake": either changes a won position to a drawn one or a drawn one to a lost one and "blunder", changes a won position to a lost one.

Who needs 25 analyses per day? Do you analyse bullet games? It is wise to analyse your lost games only and not your wins. Maybe lichess should restrict the analysis time to the actual played time.

"they actually show how many global players there are in blitz, rapid and so on (like your position being #500,000 out of 3 million players" That is available here too: go to your profile.

@lovlas Yeah I get that, and I both respect and appreciate that, equality, freedom etc. However, that doesn't mean Lichess can't improve. Since computer analyses put a lot of strain on the servers, if anything should be an extra for patrons, it should be unlimited computer analyses. Other than that, all other features should be the same for everyone. But if you ask me, I think the Lichess app should include Stockfish so that we can use our own hardware to analyze our games.

@Solal35 I'm well aware of the ratings distribution page, but it's lacking a lot of information which is available on chess.com, like just compare the two:

www.chess.com/stats/live/blitz/eliasalucard
www.chess.com/leaderboard/live

^^ Lots of details there, not shown in Lichess' ratings distribution page. For example, what's the average rating for blitz players here on Lichess? You'd have to be in exactly the 50% percentile to know the average rating here on Lichess. How many total blitz players are there on Lichess and what's your rank here among the 100 thousand players, and so on.

@Toadofsky the issue isn't really the speed of the computer analysis, but rather, that it's only available server-side. This is why we're limited to only 25 per day. Future versions of the Lichess Android app should have an option to install Stockfish so that we can do computer analyses using our own hardware, that way, removing all the load on Lichess servers, and we could have unlimited computer analyses that way too. This would totally render chess.com useless since we get unlimited, full computer analyses here for free.

@tpr yes, I really do analyze all my games, bullet included.

I agree with Danny, even in a won game you likely made mistakes. You may have won, but that was likely because your opponent didn’t notice said mistakes or did not know how to capitalise on them.

Reconnecting