Re: [computer-go] How to improve my minimax speed?

[Mark Boon]

On 17-nov-06, at 07:15, Eduardo Sabbatella wrote:


Shouldn't base the entire game play on the last move.

But looking at the last move could be an excellent
search optimisation. Indeed, I think any serious Go
program "should" look closer at the last move. ;-)



I think most of you approach the idea of proximity to the last move  
from the wrong angle. The concept to consider is stability of a  
position, not proximity in a spatial sense. Proximity is only a side- 
effect of an unstable situation in that a local answer is needed to  
make the position stable again. Statistically this is often an answer  
near the last move, but often it's also near the last move in the  
sense for example that the move defends a group next to the last move  
that was threatened by it. I'd say that the majority of the positions  
in Go are unstable and that the majority of the moves are local  
answers. Far more than the statistics show how often a move is near  
the last move in a spatial sense.


Mark

___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] How to improve my minimax speed?

[Andrés Domínguez]

2006/11/19, Mark Boon <[EMAIL PROTECTED]>:


On 17-nov-06, at 07:15, Eduardo Sabbatella wrote:

> Shouldn't base the entire game play on the last move.
>
> But looking at the last move could be an excellent
> search optimisation. Indeed, I think any serious Go
> program "should" look closer at the last move. ;-)
>

I think most of you approach the idea of proximity to the last move
from the wrong angle. The concept to consider is stability of a
position, not proximity in a spatial sense. Proximity is only a side-
effect of an unstable situation in that a local answer is needed to
make the position stable again. Statistically this is often an answer
near the last move, but often it's also near the last move in the
sense for example that the move defends a group next to the last move
that was threatened by it.


I think you are right, but proximity is easy and fast to implement while
stability is very difficult. If a program knows about stability, it understands
Go (life and death, shape, tesuji etc.). I think proximity is a good and
fast heuristic to get candidate moves.

Andrés Domínguez
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/