- Blind mode tutorial
lichess.org
Donate

Ratings Science is Hard

"In theory (Elo/Glicko), the first move advantage should be the same at every rating"

  • No. Elo/Glicko assume you get an equal portion of games with white and black.

At lowest levels the white advantage is biggest: 1 e4 e5 2 Qh5 Nc6 3 Bc4 Nf6 4 Qxf7#
That is extreme, but there are many ways to overrun black if he cannot defend well.
Danish Gambit, Evans Gambit, King's Gambit are all dubious, but effective in overrunning weaker players.

At highest level ICCF 50 days/10 moves white advantage is zero: all draws
https://www.iccf.com/event?id=100104

At top human levels classical time control the white advantage is near zero.
Gukesh won 2024 Candidates' with 2 white wins and 2 black wins.
Nepomniachtchi won 2022 Candidates' with 2 white wins and 3 black wins.
https://www.chessgames.com/perl/chess.pl?tid=114397&pid=54683

"In theory (Elo/Glicko), the first move advantage should be the same at every rating" * No. Elo/Glicko assume you get an equal portion of games with white and black. At lowest levels the white advantage is biggest: 1 e4 e5 2 Qh5 Nc6 3 Bc4 Nf6 4 Qxf7# That is extreme, but there are many ways to overrun black if he cannot defend well. Danish Gambit, Evans Gambit, King's Gambit are all dubious, but effective in overrunning weaker players. At highest level ICCF 50 days/10 moves white advantage is zero: all draws https://www.iccf.com/event?id=100104 At top human levels classical time control the white advantage is near zero. Gukesh won 2024 Candidates' with 2 white wins and 2 black wins. Nepomniachtchi won 2022 Candidates' with 2 white wins and 3 black wins. https://www.chessgames.com/perl/chess.pl?tid=114397&pid=54683

@tpr said in #2:

"In theory (Elo/Glicko), the first move advantage should be the same at every rating"

  • No. Elo/Glicko assume you get an equal portion of games with white and black.

I suppose what I mean is: in other systems (go, shogi), handicaps are offered according to the difference in player ratings, irrespective of what those ratings are; but also, I am making a simplifying assumption (which if false, I would expect there to be an abundance of evidence to refute, not just at world champion or TCEC levels).

Researchers have already shown that White has an advantage (ignoring engine or engine-assisted matches, since non-BOT Lichess players are human and we only need a rating system which works well for Lichess):
https://en.wikipedia.org/wiki/First-move_advantage_in_chess#Winning_percentages

Historian Aminson Antony offers that world championship games reflect Lichess' 55-45 split:

World Championship games from 1886 to 1990 showed 57.3% white scoring. Elite players rated 2700+ currently demonstrate 55.7% white advantage. At the World Cup level, we're seeing 57.7%, which sits right in line with historical patterns.
https://www.chess.com/blog/aminson12345/fide-world-cup-2025-complete-statistical-deep-dive-rounds-1-2

@tpr said in #2: > "In theory (Elo/Glicko), the first move advantage should be the same at every rating" > * No. Elo/Glicko assume you get an equal portion of games with white and black. I suppose what I mean is: in other systems (go, shogi), handicaps are offered according to the difference in player ratings, irrespective of what those ratings are; but also, I am making a simplifying assumption (which if false, I would expect there to be an abundance of evidence to refute, not just at world champion or TCEC levels). Researchers have already shown that White has an advantage (ignoring engine or engine-assisted matches, since non-BOT Lichess players are human and we only need a rating system which works well for Lichess): https://en.wikipedia.org/wiki/First-move_advantage_in_chess#Winning_percentages Historian Aminson Antony offers that world championship games reflect Lichess' 55-45 split: > World Championship games from 1886 to 1990 showed 57.3% white scoring. Elite players rated 2700+ currently demonstrate 55.7% white advantage. At the World Cup level, we're seeing 57.7%, which sits right in line with historical patterns. > https://www.chess.com/blog/aminson12345/fide-world-cup-2025-complete-statistical-deep-dive-rounds-1-2

#3
"World Championship games from 1886 to 1990 "

  • The old paradigm was 'draw with black, win with white. Since computers we know chess is a draw.
    Now black scores better at human top level because of engine preparation.
    It is easier to prepare for black than for white. As one GM put it: black is the new white.
#3 "World Championship games from 1886 to 1990 " * The old paradigm was 'draw with black, win with white. Since computers we know chess is a draw. Now black scores better at human top level because of engine preparation. It is easier to prepare for black than for white. As one GM put it: black is the new white.

@tpr said in #4:

Now black scores better at human top level because of engine preparation.
It is easier to prepare for black than for white. As one GM put it: black is the new white.

I lost a FIDE-rated game to an expert who opened with 1. a4 e5 2. h4. A serious game played at a classical time control where I played my best and got outplayed. Even so... if black is the new white, then white can lose a tempo with 1. a3 or similar.

@tpr said in #4: > Now black scores better at human top level because of engine preparation. > It is easier to prepare for black than for white. As one GM put it: black is the new white. I lost a FIDE-rated game to an expert who opened with 1. a4 e5 2. h4. A serious game played at a classical time control where I played my best and got outplayed. Even so... if black is the new white, then white can lose a tempo with 1. a3 or similar.

#5
Here is an example with 1 h3 and 2 a3 from the British Championship, classical time control.
https://www.chessgames.com/perl/chessgame?gid=1026336

#5 Here is an example with 1 h3 and 2 a3 from the British Championship, classical time control. https://www.chessgames.com/perl/chessgame?gid=1026336

As someone who also plays go: why is there an image of Go but no mention of it in the article or am I blind? :D

As someone who also plays go: why is there an image of Go but no mention of it in the article or am I blind? :D

Off topic (?)

1 move
1 sec
White/Black.
Did it matter?
Remember?

Who cares about rating. That was transmitted world wide, you were not even able to abort it. Didnt get a pany out of it, homeless.

Lot of fun. Breaking you, AND the game, in pieaces that you aré still looking for while I whisplashed you in the way for you to play AND loose. Thats something money cant buy, billed to you. Never forget that night.

I leave it there, the brust that I feel in my chest by only remembering that night Is too much to handle. Pure flames, upon you by me, AND only me.

CheerS. Thank you for everything (Thats upon to come.
My children Will fly above you!! Jaja Move sucker move!!

I leave it here. Cant stop. You know why, Is not my fault. But Sorry (?) See u, see mines, soons. Hajaj

Off topic (?) 1 move 1 sec White/Black. Did it matter? Remember? Who cares about rating. That was transmitted world wide, you were not even able to abort it. Didnt get a pany out of it, homeless. Lot of fun. Breaking you, AND the game, in pieaces that you aré still looking for while I whisplashed you in the way for you to play AND loose. Thats something money cant buy, billed to you. Never forget that night. I leave it there, the brust that I feel in my chest by only remembering that night Is too much to handle. Pure flames, upon you by me, AND only me. CheerS. Thank you for everything (Thats upon to come. My children Will fly above you!! Jaja Move sucker move!! I leave it here. Cant stop. You know why, Is not my fault. But Sorry (?) See u, see mines, soons. Hajaj

@ralgrado said in #7:

As someone who also plays go: why is there an image of Go but no mention of it in the article or am I blind? :D

That's certainly a fair question. I usually spend many minutes searching for a good thumbnail image after writing the blog post, and then I remembered that go has komi handicap for the first move.

@ralgrado said in #7: > As someone who also plays go: why is there an image of Go but no mention of it in the article or am I blind? :D That's certainly a fair question. I usually spend many minutes searching for a good thumbnail image after writing the blog post, and then I remembered that go has komi handicap for the first move.

It's pretty ironic that you didn't credit xkcd/Randall Munroe for the comics.

As for the topic: if you looked at the data and that's what the data said, then I have to admit I can't argue; but the format of this blog post probably isn't the clearest way to present an in-depth data analysis, and I think @tpr has a good point.

In other words: if all you did was consider 1500s in this analysis, then you didn't really do your due diligence, because the assumption that it's the same at all rating levels is far from obviously true. If you can point to Glickman's work and say "he proved/justified the assumption" then that's another matter. But it's not immediately clear to me from reading the blog post that either of those previous "if" conditionals are met, meaning that in total, this blog post has raised more questions (for me) than it has answered.

I guess that means I agree with the title!

It's pretty ironic that you didn't credit xkcd/Randall Munroe for the comics. As for the topic: if you looked at the data and that's what the data said, then I have to admit I can't argue; but the format of this blog post probably isn't the clearest way to present an in-depth data analysis, and I think @tpr has a good point. In other words: if _all_ you did was consider 1500s in this analysis, then you didn't really do your due diligence, because the assumption that it's the same at all rating levels is far from obviously true. If you can point to Glickman's work and say "he proved/justified the assumption" then that's another matter. But it's not immediately clear to me from reading the blog post that either of those previous "if" conditionals are met, meaning that in total, this blog post has raised more questions (for me) than it has answered. I guess that means I agree with the title!