sai training site

[l1t1]

http://sai.unich.it/

[l1t1]

the lz nets of samilar strength is about 4 times training games of sai

[dbosst]

the lz nets of samilar strength is about 4 times training games of sai

So it will take at least (almost definitely more so) ~5 million training games to reach LZ's current strength?

[l1t1]

it should be 8 times, now 150k training games. vs lz 24 68%

[nemja]

LZ had a lot of early bugs, the differences may not be constant all the way.

[Vandertic]

I believe SAI will get superhuman in a few months, with the support of the community. Anyhow we don't presume that it will ever be as strong as the latest LZ networks. That's not its purpose.

The aim of SAI is to get rid of the suboptimal moves LZ or AG may play when they are too far ahead of their opponent (i.e. almost always against humans) or when they are far behind (i.e. at the beginning of a game with handicap).

When SAI will be superhuman, it could show really how much handicap (in points or maybe in stones) it could give to a strong player. Or how many points difference it could score in a fair game.
In endgame, it may be able to show unknown smart plays or exhibit optimal choices in difficult situations.

This is presently from very difficult to impossible under the LZ/AG structure, because you would need to tailor positions in which the perfection of endgame moves really would matter for winning, and not just for score.

[Glrr]

Sai's progress would be even faster with the same number of visits as LZ (sai uses only 800 visits right now).