Many Faces has the same problem. Chinese rules, it thinks O wins about 70%.
-Original Message-
From: computer-go-boun...@computer-go.org [mailto:computer-go-
boun...@computer-go.org] On Behalf Of Brian Sheppard
Sent: Friday, July 10, 2009 7:05 PM
To: computer-go@computer-go.org
Valkyria correctly behaves as if this is seki. It actually does not
now it is seki explicitely, but it prunes all moves that would break
the seki.
The principle to do this as simple as possible is thus not actually to
identify seki in general but to find simple rules that prunes bad
One of the difficult questions is if (or better how)
dynamic komi can be used to improve the strength of
MCTS go programs in handicap games (both cases being
interesting: computer on strong side - and -
computer on weak side).
Especially, there are several normal go players
(non-programmers) who
Hello,
I'm thinking about wether it's better to keep a list of stones for
each group in the board state or not.
I'm keeping a linked list of liberties like Lukasz suggested and it is
useful in many cases, but the only case that access to all stones in a
group is handy is when removing
4. Keep a singly linked list of stones for each group and keep an
additional pointer for each group to the last element of the list. This is
what I do in Many Faces.
David
-Original Message-
From: computer-go-boun...@computer-go.org [mailto:computer-go-
boun...@computer-go.org] On
On Sat, 2009-07-11 at 08:57 -0700, David Fotland wrote:
4. Keep a singly linked list of stones for each group and keep an
additional pointer for each group to the last element of the list. This is
what I do in Many Faces.
5) Use a bitmap. This costs a bit more memory (if one bitmap is
Isn't 4. similar to doubly linked lists? You have to keep almost as
many pointers as there are points on the board at most. How do you
effectively store the pointers to only use as few as possible?
I don't see how 5) is good for removing groups. Are there other uses
for the bitmaps?
Am
Yes, 4 is very similar to doubly linked lists in memory size. I think it's
a little faster. To save memory, I don't use pointers. I use 2-byte
indexes into arrays.
David
-Original Message-
From: computer-go-boun...@computer-go.org [mailto:computer-go-
boun...@computer-go.org] On
I think we should open up to other ideas, not just dynamic komi
modification. In fact that has not proved to be a very fruitful technique
and I don't understand the fascination with it.
First we identify what it is we are trying to accomplish. You mentioned
improving the strength of MCTS go
The CGOS page should have a list of the GTP commands required by CGOS.
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
If you are in a lost position, good play is play that maximizes
the probability of a turnaround, which is quite different depending
on how far behind you are, and for what reason.
If the status of all the major groups is solid, then concentrating
on tactics which can gain a few points reliably
The go-playing literature offers a bit of advice: when ahead, make moves which
simplify the game and preserve your advantage. When behind, take some risks to
grab more than you are entitled to - but not too many. Computer programs seem
bizarre in this regard, they tend to play quite
On Sat, Jul 11, 2009 at 4:54 PM, Dave Dyer dd...@real-me.net wrote:
If you are in a lost position, good play is play that maximizes
the probability of a turnaround, which is quite different depending
on how far behind you are, and for what reason.
What maximizes the probability of a
i think that the rationale behind variable komi is intuitive:
good players can handicap one another more effectively
with komi than with handicap stones, because it's more
fine-grained.
this is likely what is leading to the idea that computers
playing handicap games could use this to their
I have thought about this kind of thing a bit, but it's hard to
formulate a general solution. What happens is when you prohibit
certain 'bad' moves is that you slant the result in favour of the side
with more 'bad' moves. This gets to be a problem in situations where
there are few moves to
What do use for your counters? 32 bit numbers max out at 4 billion,
and you're already beyond that.
Is it possible to generate an SGF file showing the dominant variations
with the number of wins and losses? It'd be interesting to see what
the bot considers to be the best sequences are...
16 matches
Mail list logo
