Actually, trying to beat a bot with any special strategy doesn't make
a lot of sense.
Just play your best go and be happy that you have an interesting opponent.
I know that there is a track record of people initially losing against a
bot, and beating it afterwards.
But that is mainly due to the respect the bot has earned for the things
it does well, attacking and moyo building.
Pathological weaknesses should be of more interest to the programmers
than to anybody else.
The semeai problem is a tough nut for 2 reasons:
1. As David Fotland has stated, if one side has to play in a certain
order and the other side can play in any order, playouts are strongly
biased towards the side that has many possible moves.
2. Semeais are relatively simple to solve in isolation, but not when
irrelevant plays in other parts of the board are intermixed with the
analysis.
This is very hard to solve, because in most other contexts global search
is an important mc strength, not a weakness.
Stefan
Am 15.09.2010 05:27, schrieb Jeff Nowakowski:
On 09/14/2010 04:36 PM, terry mcintyre wrote:
From my observations of human-versus-bot games, a winning strategy
against bots
seems to be:
Create several capturing races, even if you lose all of them.
Is there an established, reliable way to create capturing races
against bots?
_______________________________________________
Computer-go mailing list
Computer-go@dvandva.org
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
_______________________________________________
Computer-go mailing list
Computer-go@dvandva.org
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
