[gnugo-devel] SlugGo v.s. Many Faces

[David G Doshay]

Hello all,

Over the last year I have mentioned a few times that I am working on a 
Go project. We call our code SlugGo, partly because the banana slug is 
the mascot of UCSC where the work is being done, and partly because the 
program runs very slowly in spite of the fact that it is spread over a 
cluster of the latest G5 Macs.

SlugGo starts with GNU Go and adds parallel whole board lookahead. 
SlugGo uses GNU Go to generate candidate moves from each player's 
perspective and adds to the GNU Go move value additional values 
depending upon the whole board positions at the end of the lookahead to 
determine which of the candidate moves to select. The primary reason 
the code is so very slow is that we have found a lookahead depth of 16 
works very well, and it thus takes a bit more than 16 times as long as 
GNU Go to settle on its move.

We have been encouraged by our success against the GNU Go base code, 
where we can win about 70% of the games when giving GNU Go a 5 stone 
handicap, but were worried that our success was primarily the result of 
our correctly guessing the responses to our move choices because they 
are all generated by GNU Go, that is to say that we know exactly how 
our opponent thinks.

We are now confident that we are indeed much stronger than GNU Go 
because of our results against Many Faces of Go.

In a 15 game set against Many faces, with Many Faces playing Black and 
no komi, the results are:

Black wins by 33
        White wins by 126
        White wins by 59
        White wins by 67
Black wins by 15
        White wins by 73
        White wins by 40
        White wins by 17
        White wins by 25
A very complicated endgame board where perhaps White wins by 76, but if 
play were to resume again perhaps black could pull off a win by 
something like 20. I have not worked this one out completely.
Black wins by 75
        White wins by 58
        White wins by 21
        White wins by 125

The sgf files for these games can be downloaded from:

        http://hsrf-mc2.cse.ucsc.edu/~cowboy/mfgo_sluggo.tar.gz

We also have 10 more game results, also with MFG playing Black and no 
komi:

        White wins by 24
        White wins by 2
        White wins by 42
        White wins by 54
        White wins by 41
        White wins by 1
Black wins by 4
        White wins by 32
        White wins by 53
Black wins by 21

but we have not posted these sgf files yet.

So, SlugGo wins about 20 out of 25 from Many Faces. We will continue 
this contest against Many Faces until we have 100 games.

We will continue to make our results available as we have them.

Cheers,
David