- Blind mode tutorial
lichess.org
Donate

If Stockfish had a FIDE rating what would it be?

#1
<Comment deleted by user>
#3

@fiskaren Of course, most of the top programs (rated 3200+ on CCRL) must be extremely strong, but there is almost no testing of programs against top players in a classical tournament setting. So in that sense the actual numbers like 3500 are meaningless in terms of FIDE rating. I think they should have regular classical tournaments that include a couple of engines so that we will at least have some idea. For example, as of today Magnus is 2834 (classical), and SF is 3549 (at 40 moves in 15 minutes, so relative to humans would probably be even more at classical time controls). A rating difference of 715 translates to an expected score of 99.38%. For 3200 engine rating, we get chances of about 81% (engine win), 1% (Magnus win), 18% draw. So especially against not so strong engines (like a 3200 engine), it should not be hard to test if the CCRL ratings translate well to FIDE ratings.

@fiskaren Of course, most of the top programs (rated 3200+ on CCRL) must be extremely strong, but there is almost no testing of programs against top players in a classical tournament setting. So in that sense the actual numbers like 3500 are meaningless in terms of FIDE rating. I think they should have regular classical tournaments that include a couple of engines so that we will at least have some idea. For example, as of today Magnus is 2834 (classical), and SF is 3549 (at 40 moves in 15 minutes, so relative to humans would probably be even more at classical time controls). A rating difference of 715 translates to an expected score of 99.38%. For 3200 engine rating, we get chances of about 81% (engine win), 1% (Magnus win), 18% draw. So especially against not so strong engines (like a 3200 engine), it should not be hard to test if the CCRL ratings translate well to FIDE ratings.
#4

You are just too late to the party. The days when grandmasters were willing to play against machines are over and will never come back. It is just a question of how much of a victim the human is.
We also had matches of stockfish against world computer chess champions of the past and they ended also very humiliating, like 1000:0, don't have the references handy.
Them Komodo team around Larry Kaufman experimented with game odds against grandmasters, and i think two pawns up in the opening still looses the game most of the time, chances depending on which two pawns are missing.

You are just too late to the party. The days when grandmasters were willing to play against machines are over and will never come back. It is just a question of how much of a victim the human is. We also had matches of stockfish against world computer chess champions of the past and they ended also very humiliating, like 1000:0, don't have the references handy. Them Komodo team around Larry Kaufman experimented with game odds against grandmasters, and i think two pawns up in the opening still looses the game most of the time, chances depending on which two pawns are missing.

This topic has been archived and can no longer be replied to.