# AlphaGo: Where's AlphaChess? **URL:** \ **Category:** The Game Room\ **Created:** [October 19, 2017, 8:22pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319 "2017-10-19T20:22:34Z")\ **Posts on this page:** 13\ **Page:** 3
**Author:** ![pulykamell](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/pulykamell/32/3166_2.png) [@pulykamell](https://boards.straightdope.com/u/pulykamell)\ **Post date:** [December 7, 2017, 4:38pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/41 "2017-12-07T16:38:14Z")
> [@borschevsky](#): > > I’ve read a bit more about this, and some people are pointing out that the hardware used for the Stockfish match may not really have been fair. AlphaZero is running on high-powered custom hardware, while Stockfish wasn’t, and was given a relatively small hash table size and no access to tablebases. So maybe the result isn’t _quite_ so significant, in terms of the strength of the AlphaZero engine itself. If you put Stockfish itself on similarly-powered hardware, it would probably beat the testing version of Stockfish handily as well. Yeah, I saw that, but even if Stockfish were somewhat hobbled, I find AlphaZero Chess an impressive result for a self-learning system with no access to any outside inputs other than the chess rulebook. ---
**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\ **Post date:** [December 8, 2017, 3:35am UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/42 "2017-12-08T03:35:46Z")
Another thought: Stockfish was programmed by human chess masters, and designed to deal with what humans expect to see. Novelty, itself, is a weapon against such. It wouldn’t be the first time something like this has happened: POWs in Vietnam found themselves playing a lot of chess, because it was about the only way they could spend their time. None of them was particularly good at the start, and they knew little about conventional theory, but they ended up learning a lot just by practice. And when they were freed, it took a while for the conventional chess world to adapt to their unconventional techniques. Well, here again we have someone becoming highly proficient, without contact with conventional theory. Maybe the conventional wisdom (as implemented in Stockfish) just doesn’t know how to react, again. Well, not “just”: That obviously can’t explain all of a hundred-game lossless streak, but it might be part of it. The logical test to perform would be to take multiple copies of AlphaZero, let them learn chess independently, and then play them against each other to get relative rankings. Would they all be about equally good, or would some of them have developed ideas that surprise even their siblings? ---
**Author:** ![N9IWP](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/n9iwp/32/3154_2.png) [@N9IWP](https://boards.straightdope.com/u/N9IWP)\ **Post date:** [December 8, 2017, 12:01pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/43 "2017-12-08T12:01:23Z")
Related ArsTechnica article: [DeepMind AI needs mere 4 hours of self-training to become a chess overlord | Ars Technica](https://arstechnica.com/gaming/2017/12/deepmind-ai-needs-mere-4-hours-of-self-training-to-become-a-chess-overlord/) Chess 24article: [https://chess24.com/en/read/news/deepmind-s-alphazero-crushes-chess](https://chess24.com/en/read/news/deepmind-s-alphazero-crushes-chess) > [@](#): > > Meanwhile, though, it’s gratifying to see that the computer has justified 100s of years of chess development, since the program, entirely by itself, has ended up playing some of the best known human openings… The graphs are fascinating to study, since you can see how certain openings became popular in the algorithm’s training games – such as the French Defence and the Caro-Kann – before dropping off in popularity as its strength increased. Brian ---
**Author:** ![BigAppleBucky](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/bigapplebucky/32/178_2.png) [@BigAppleBucky](https://boards.straightdope.com/u/BigAppleBucky)\ **Post date:** [January 17, 2018, 1:17pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/44 "2018-01-17T13:17:46Z")
I’m going by memory so these numbers aren’t precise, but they’re close. Saw them on a number of YouTube videos. The top Elo rating for a human is about 2,700 or 2,800. Stockfish, the engine chess champion had a rating about 3,300. Based on games so far, it is estimated AlphaZero’s rating is about 4,100 which is astonding to me. When I played I was about a 1,500 player. Frequently played an A rated player (1,800) and lost nearly every time. I probably won 3 or 4 times in a 100 games. He, in turn, once beat a master level player. Just that once, but he was very proud of that game. A gap of 800 Elo points between Stockfish and AlphaZero would be huge. ---
**Author:** ![Snarky\_Kong](https://avatars.discourse-cdn.com/v4/letter/s/a183cd/32.png) [@Snarky\_Kong](https://boards.straightdope.com/u/Snarky_Kong)\ **Post date:** [January 17, 2018, 1:47pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/45 "2018-01-17T13:47:09Z")
> [@BeepKillBeep](#): > > I hope you can find it, I’d like to read it. So, turns out I was conflating two papers. [https://arxiv.org/pdf/1710.05941.pdf](https://arxiv.org/pdf/1710.05941.pdf) Has the activation function search and introduces the swish function which has got a little attention [https://arxiv.org/pdf/1709.07417.pdf](https://arxiv.org/pdf/1709.07417.pdf) Has a search for gradient descent update rules. ---
**Author:** ![borschevsky](https://avatars.discourse-cdn.com/v4/letter/b/97f17d/32.png) [@borschevsky](https://boards.straightdope.com/u/borschevsky)\ **Post date:** [January 17, 2018, 4:02pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/46 "2018-01-17T16:02:28Z")
> [@BigAppleBucky](#): > > The top Elo rating for a human is about 2,700 or 2,800. Stockfish, the engine chess champion had a rating about 3,300. Based on games so far, it is estimated AlphaZero’s rating is about 4,100 which is astonding to me. That number for AlphaZero’s rating doesn’t make sense. Since it scored 64% against Stockfish, its rating should be 100 points higher than Stockfish’s. I’ve seen estimates for the maximum possible rating at about 3600. If you imagine a perfect chess-playing entity, would Magnus Carlsen draw one out of every 50 games against it, and lose the other 49? If so, its rating would be about 3600. ---
**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\ **Post date:** [January 17, 2018, 4:53pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/47 "2018-01-17T16:53:31Z")
That depends on what perfect play leads to, doesn’t it? If, as seems plausible, it’s possible for either side to force a draw in chess, then one would expect the proportion of draws to increase as skill level on both sides increases, and that past some skill level, a player would very seldom lose, even against a far more skilled opponent. Of course, this might mean that the standard definition of the rating system fails to adequately discriminate between players above some skill level: That is, maybe the maximum possible rating really is 3600, but a player with a rating of 3599 might actually be far, far better than a player with a rating of 3598. ---
**Author:** ![DPRK](https://avatars.discourse-cdn.com/v4/letter/d/4491bb/32.png) [@DPRK](https://boards.straightdope.com/u/DPRK)\ **Post date:** [January 17, 2018, 5:04pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/48 "2018-01-17T17:04:59Z")
> [@Snarky\_Kong](#): > > So, turns out I was conflating two papers. > > [https://arxiv.org/pdf/1710.05941.pdf](https://arxiv.org/pdf/1710.05941.pdf) Has the activation function search and introduces the swish function which has got a little attention > > [https://arxiv.org/pdf/1709.07417.pdf](https://arxiv.org/pdf/1709.07417.pdf) Has a search for gradient descent update rules. Thanks! ---
**Author:** ![DPRK](https://avatars.discourse-cdn.com/v4/letter/d/4491bb/32.png) [@DPRK](https://boards.straightdope.com/u/DPRK)\ **Post date:** [January 17, 2018, 5:14pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/49 "2018-01-17T17:14:38Z")
> [@Chronos](#): > > That depends on what perfect play leads to, doesn’t it? If, as seems plausible, it’s possible for either side to force a draw in chess, then one would expect the proportion of draws to increase as skill level on both sides increases, and that past some skill level, a player would very seldom lose, even against a far more skilled opponent. Of course, this might mean that the standard definition of the rating system fails to adequately discriminate between players above some skill level: That is, maybe the maximum possible rating really is 3600, but a player with a rating of 3599 might actually be far, far better than a player with a rating of 3598. I suppose you’re right: only the difference in rating is important in Elo’s system, not the absolute score. Also, only results matter, so a much stronger player who is able to win 1% of the time will not have a much greater rating. Chess is simple enough for these superhuman opponents that they will all have a similar (high) rating, just as they will for Tic-Tac-Toe. What the maximum rating is will depend on previous calibration, then. ---
**Author:** ![borschevsky](https://avatars.discourse-cdn.com/v4/letter/b/97f17d/32.png) [@borschevsky](https://boards.straightdope.com/u/borschevsky)\ **Post date:** [January 17, 2018, 5:15pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/50 "2018-01-17T17:15:05Z")
> [@Chronos](#): > > That is, maybe the maximum possible rating really is 3600, but a player with a rating of 3599 might actually be far, far better than a player with a rating of 3598. Far better by what metric? A 3599 scores only 50.1% or so against a 3598. If chess is a draw, then it seems plausible that Carlsen could draw 1 out of every 50 games against a perfect player. If chess is a win for white (or black), then Carlsen would lose all his black games, but plausibly could draw 1 out of 25 white games. ---
**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\ **Post date:** [January 17, 2018, 7:25pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/51 "2018-01-17T19:25:25Z")
I guess what I’m getting at is that, at sufficiently-high levels, the assumptions behind the Elo rating system probably break down. One can envision two extremely powerful players who _almost_ always draw against each other, but such that A can beat B one game out of 100, while B can’t beat A even one time in a billion. If B only makes a game-losing error 1% of the time, but A only makes a game-losing error less than one game in a billion, then in some sense A is ten million times better… but their ratings will be very close. ---
**Author:** ![BigAppleBucky](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/bigapplebucky/32/178_2.png) [@BigAppleBucky](https://boards.straightdope.com/u/BigAppleBucky)\ **Post date:** [January 19, 2018, 5:14pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/52 "2018-01-19T17:14:51Z")
> [@borschevsky](#): > > That number for AlphaZero’s rating doesn’t make sense. Since it scored 64% against Stockfish, its rating should be 100 points higher than Stockfish’s. > > I’ve seen estimates for the maximum possible rating at about 3600. If you imagine a perfect chess-playing entity, would Magnus Carlsen draw one out of every 50 games against it, and lose the other 49? If so, its rating would be about 3600. [![]( " - YouTube") ](https://www.youtube.com/watch?v=eN7BMWl_mpw&t=1s) In these games the engines were given the first few moves of human concieved openings. Not sure what the engines would do without those openings. Maybe white has a forced win or maybe with best play the games are always draws. > [@](#): > > 1200 games of 12 openings, 100 games per opening were played by AlphaZero & Stockfish 8 that are in the paper released by the developers. 10 games from the paper are in current videos. I haven’t seen any one mention anything about the other games??? So I did this video & came up with a rating from the data, & also expressed some obvious implications of it. > > Both programs played back & white 50 games in each opening. The results for Alpha Zero were: > > As white; 242 wins, 353 draws, 5 losses. > As black; 48 wins, 533 draws, 19 losses. > For a total of 290 wins, 886 draws, 24 losses. > > 23% wins, 73.83% draws, 2% losses. So it did lose to Stockfish, why nobody is mentioning it I don’t know but I’m assuming investment reasons & hype. [![]( " - YouTube") ](https://www.youtube.com/watch?v=M1Dva2KCmBw&t=616s) AlphaZero was run on a super computer platform, apparently Stockfish was not. That might have been extremely important. ---
**Author:** ![borschevsky](https://avatars.discourse-cdn.com/v4/letter/b/97f17d/32.png) [@borschevsky](https://boards.straightdope.com/u/borschevsky)\ **Post date:** [January 19, 2018, 5:40pm UTC](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319/53 "2018-01-19T17:40:31Z")
> [@BigAppleBucky](#): > > [https://www.youtube.com/watch?v=eN7BMWl\_mpw&t=1s](https://www.youtube.com/watch?v=eN7BMWl_mpw&t=1s) He writes the following in the video description: > [@](#): > > For the rating compared to Stockfish which is around 3400, AlphaZero won 23% of the games & lost 2%, for a 21% strength over Stockfish, which is 700 points & a total rating of 4100+. He’s apparently taking that 21% and saying that 3400 + (3400 \* 0.21) = 4100+. That is not remotely close to how the ratings work. Winning 23%, losing 2%, and drawing 75% against a 3400 gives you a rating of 3474. [Previous page](https://boards.straightdope.com/t/alphago-wheres-alphachess/799319.md?page=2)