There's no way the world's greatest chess-playing entity is breaking any sweat here. After all, it mastered chess in a mere four hours of practice. Answer (1 of 6): It’s beautiful and will create waves of change in the way chess community approach the game. It's not even a sporting match for a genuine fight, Alpha Zero needs to start the game, like, a knight down. *Stockfish plays great and never blunders, but it never has Alpha Zero even remotely in trouble. ![]() Watching these Stockfish-Alpha Zero games is like watching a powerful bulldozer utterly baffled by a drone with an overhead view. It's doing what neurons do, it seems to resolves situations visually, like the nerves at the back of an eyeball. Alpha Zero just doesn't number-crunch the situations on the board. *Commentators are describing the neural net's attack style as "strangling" or "smothering," and, although I'm plenty far from a chess expert, I can see that. Alpha Zero locks Stockfish into a position where every possible move it can make is either lousy or fatal. It's Google's deep learner neural net intuitively smothering the most powerful chess engine in the world. Update: with wins in the last four TCEC championships, Stockfish is probably the strongest engine right now, up to small sample size effects (see first caveat above).*If you've got fifteen minutes to spare for the royal game of chess, you should watch this. They don't perceive it as a challenge, however, because they're not a chess engine company and don't view having the best engine as a big deal. For the foreseeable future, this situation is likely to remain the case.Įdit: Deepmind acknowledges that AlphaZero would probably lose to the latest versions of Stockfish and Leela. For the question "who is the strongest engine right now?", the situation is fluid there are several engines that could all claim to be #1 at any one time. Leela is at least competitive with Stockfish, and there's a good chance it's stronger (right now) as well. Tl dr: AlphaZero is probably inferior to Stockfish 10, and even more so to the latest versions of Stockfish. You might be interested in the ongoing Season 18 of the Top Chess Engine Championship, live games are available here. Plus, as of time of writing, all three of Leela / AllieStein / Stockfish are behind Stoofvlees in the qualification tournament for the superfinal. Like Leela, AllieStein is a neural network engine unlike Leela, AllieStein includes human knowledge. In fact just two seasons ago, Leela actually failed to qualify for the superfinal, losing out to AllieStein. Stockfish is not Leela's only competitor.So while Leela was (probably) stronger two months ago, it might not be the strongest engine right now. As of time of writing, Leela has kinda stalled while a new net is being trained. Still, we can't play ten thousand games at long time controls - that takes too long - so we have to be satisfied with 100. These games are played at bullet time controls, which isn't really chess as us humans understand it. Engine developers regularly need tens of thousands of games to conclusively say that one engine is stronger than the other. In the world of computer chess, 100 games is an incredibly small sample size.Indications are it's competitive with Stockfish, and (for now at least) stronger: it won the most recent version of the Top Chess Engine Championship by five games - an elo difference of about 15. It incorporates many new innovations not in the original paper, and therefore should be stronger than AlphaZero. After winning 25 games of chess versus Stockfish 8 starting as white, with first-mover advantage, a further three starting with black and drawing a further 72 games, AlphaZero also learned shogi. Having said that, Leela Chess Zero is an engine based on AlphaZero techniques. ![]() You ask about Stockfish 10 (not latest Stockfish), which still beats Stockfish 8 by 100 elo, so it should be capable of beating AlphaZero as well. ![]() engine A can beat engine B, which beats engine C, but C could still beat A in a match) however, the differences are so large that it would be incredible if AlphaZero is capable of beating the latest versions of Stockfish. Stockfish started out suggesting some moves, but after 6 minutes on my computer (corresponds to 7. It's true that elo isn't transitive (i.e. The latest versions of Stockfish are capable of beating Stockfish 8 by about 150 elo (see also this). That's because AlphaZero beat Stockfish 8 by +155 = 839 -6, which is an elo difference of about 50. Still, if AlphaZero hasn't improved since it was unveiled, it will likely lose to the latest version of Stockfish. we don't have games between it and the latest versions of Stockfish. We can't say for sure since AlphaZero is a private engine, i.e.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |