Right now, there’s a player lurking in the depths of the online Go scene that is laying waste to some of the best players in the world. It’s called Master, and nobody knows who it is.
Update: The identity of the mystery account has finally been revealed – you can read all about it here.
The account is simply called “Master”, and since the start of the new year it has made a habit out of trashing some of the world’s best Go professionals. It’s already beaten Ke Jie twice, who is currently the highest ranked Go player in the world. AlphaGo, incidentally, is #2.
Not that the ranking stopped him from being battered, mind you. A European professional Go player, Ali Jabarin, wrote on Facebook that Ke Jie was “a bit shocked … just repeating ‘it’s too strong’”. Jabarin wasn’t sure whether the player was AlphaGo or not, but he was certain that an AI was behind the mystery account.
By January 3, the number of probably-but-we-can’t-officially-say AI sanctioned beatings had risen to 41-zip. There’s a few signs that it might not be an all-AI account, though. Jabarin received a polite message on New Year’s declining a match, and a post appeared offering around $US14,000 to any professional player who could beat it.
Given AlphaGo’s beatdown of Lee Sedol last year, the initial speculation blamed Google for the carnage. But given the way Google so prominently publicised AlphaGo’s matches, it’s unlikely they would just unleash a champion-beating AI on the world of online Go without so much as a post. And even if they were to do so, it’s already been noted that Google already has two accounts reserved on the most prominent Chinese and Korean online Go servers.
The suspicion, then, is that another autonomous Go legend has appeared. While AlphaGo was getting all the praise last year, another company launched a deep learning Go bot with the support of the Japanese Go association.
The makers estimated it was about 500 ELO points behind AlphaGo at the time – although the bot is also a known quantity, plays under a different alias and reportedly has a different playstyle. South Korea’s amateur Go association has been developing an AI as well, and it wouldn’t be surprising if the developers of that unleashed their efforts online just to test the waters. Tencent has been developing a Chinese Go AI too, as has Acer.
But this is all purely speculation, as there’s been no official confirmation as to the mystery player wrecking online Go. The only thing anybody knows for sure is that the world’s best Go players have been getting slapped around by something. Possibly someone, but probably something. If you’d like to review the slaughter, I’ve embedded one of the matches against Ke Jie above. There’s also a good post on a German Go site with embeds of Master’s first 41 games.
Update: Master’s winning streak has finally been ended – although not through entirely legitimate means. Master racked up 51 straight wins, only for the internet connection of professional Chen Yaoye to time out. The timeout meant that the 52nd game on Master’s record will be recorded as a tie, therefore breaking the winning run.
Comments
24 responses to “Someone Is Destroying Online Go, And Nobody Knows Who It Is”
Oooh, it’s like one of those urban myths, like in my Japanese animes!
Weeb
It’s me.
The Machine is out there!
Ok what is online GO? Excuse me for my sheltered life.
The online community for the ancient board game Go, a one-on-one strategy game where you place stones to capture more territory than your opponent.
Copy that thanks, and people make “bots” for it and the community is cool with this?
I assume they treat it much the same way as you would making bots for chess. The rules are fixed, so being robotic doesn’t confer much advantage in the way that a bot for an MMO might be an advantage of having infinite stamina. People beat chess bots all the time because they come with different disadvantages and can be exploited. I assume the same is true of Go-bots.
Ah right yeah never thought to look at it that way. Learning all the time.
Being AI offers huge advantages. A computer can look hundreds of moves ahead and statistically work out the best possible move. IBM’s AI beat out the best chess player in the world years ago and they’ve only gotten better since then with the explosion of processing power now available. Deep Blue played chess in 1997 and could evaluate 200 million positions a second. Unbeatable then, even more unbeatable now.
Deep Blue was beaten several times though. Computers can look ahead well but they lack intuition and can’t read bluffs well. That’s what makes it an interesting contest, because it pits raw calculation against intuition and the former doesn’t always come out on top.
Except the number of possible moves in a game of Go, is approximately 1×10^171, therefore a bot cannot brute force a winning strategy as it can for chess. Programming a good Go bot is an exponentially more dificult task than a chess bot.
Crazy and mysterious @alexwalker. Thanks for sharing. Also: Skynet.
Skynet?
Clearly Sai is back!!
I understand this reference and it pleases me
I love that manga. Kudos good sir!
ooh Right out of Hikaru no Go….It’s Sai!!!
lunatic-handsup-Aliens.png
wouldnt it be funny if it was actually watson or deepmind legit
Great now I gots watch Hikaru again
It’s AlphaGo
https://twitter.com/demishassabis/status/816660463282954240
Well that’s disappointing.
Cool to see the game I uploaded here. The same game was also analyzed here by a 6d player, which sparked some more discussion: https://gokibitz.com/kifu/S16_4wYrg