Hacker News Reader

AlphaZero was playing Stockfish 8. This is done using Stockfish dev which is more than 110 ELO higher than Stockfish 8 (see https://github.com/glinscott/fishtest/wiki/Regression-Tests) and higher than what was reported for AlphaZero.

Check out the Leela project, which is trying to reproduce and improve the AlphaZero project to also beat Stockfish dev: http://lczero.org/ (at this point they are about equal with Stockfish 9)

View on HN

A 100 ELO difference is about a player twice as stronger. With a 100 ELO difference, Stockfish would have won some games.

28Win 72Draw and O Loss in my opinion reflects at least 300 ELO (8 times stronger).

Elo is not a matter of opinion. There's a formula for it, and 28 wins, 72 draws and 0 losses represents almost exactly a 100 point Elo difference.

https://www.3dkingdoms.com/chess/elo.htm

You are right, thank you for the pointer.

I guess that the spectacular 28-0 was somehow confusing my ELO opinion engine by concealing the fact that 72 draws are a lot more difficult to achieve than 12 draws (28-12-0 would have been a 300 ELO difference).

The A0 vs Stockfish 8 match also didn't use endgame tablebases.

Reply on news.ycombinator.com