lichess.org
Donate

Finally a decent Go server

#20 wasn't aware of WBaduk but I'll just note for newcomers that Kiseido is a client for the famous Korean Go Server (KGS). It is the biggest Go server and also the most skilled. Just beware that the skill level is MUCH higher than on Western-friendly sites. It's like the Stockfish level 8 of go servers. Of course if you want that, go right ahead. :)
I just recently discovered this server too and I love. still learning the very basics of the game but it sure is hella interesting
<Comment deleted by user>
THE BEGINNING OF THE END !?

http://www.bbc.com/news/technology-35420579
www.youtube.com/watch?v=PNl7AjzfsYU

That is a real breakthrough indeed and I am not sure whether to be happy or sad about it. It was believed to be next to impossible.

10 to the power of 171 possible positions and on top of it very subtle and complex concepts that are impossible to be put into a formula ( or "pawn units for that matter " ) like aji ( latent possibilities ), amashi ( inducing strategy ) , atsumi ( thickness ) katachi ( shape ), gorikatachi ( overconcentration ) ko, double ko, triple ko and and and.

Even though the European Champion is still classes below a real top Asian Professional 9-Dan, he is also classes above the average amateur .

In Elo terms he would probably be in the range of 2300+
( http://senseis.xmp.net/?FIDETitlesAndEGFGoRatings )

One of my reasonings for giving up concentrating on Chess to play Go many years ago was that the mysterious beauty of Chess was taken away by the brutal calculation reality of the engines... and now ?

The sunny side of it of course is that now both games will become equally sober and demystified... but surely still worth our time ...

Cheers M.

@26 Absolutely, though in another article it was said to be even higher, like 10 to the power of 700 (!) but I guess there will always remain some uncertainty of that because of the theoretical possibility of "ko" recapturing endlessly and the fact that stones or groups are captured and the empty spaces can be refilled later, just to be captured again, though that doesn't make a lot of sense strategically. Also there is no 50 move rules (and no draws are there also at least with modern rules and 1/2 komi )

It won't be long and the GO server sites will need to implement engine detection software...

But since you mention 19x19... there is of course a possibility to enlarge the board even more, just like there have been ideas to add pieces to the chess set ( or rearrange like Fischer Random )
but... it wont be the same and the fact that this AI guy mentioned that the engine will play against itself and LEARN ... will probably mean its just a question of time....

Interesting times....
To whom it may be of interest :-)

DeepMind AlphaGo vs Lee Sedol

Lee Sedol will play a five game match against Google DeepMind's AlphaGo computer program in March 2016.

After defeating Fan Hui 2p with a 5-0 score, Google DeepMind chose to challenge Lee Sedol because of his record as the best Go player in the world over the last decade.

The games will be even, with $1 million USD in prize money for the winner.

According to the Korean Baduk Association, the match will take place from March 8 to 15, 2016, in Seoul, Korea.
I basically don’t understand anything but for some reason I found the first 4 games riveting!

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