RE: computer-go: Applying Moore's Law to Computer Go

[David Fotland <fotland@xxxxxxxxxxxxxxx>]

Sorry, I should have said move creates a single stone group with
one liberty and captured exactly one stone.

David

At 03:49 PM 11/24/99 +0800, you wrote:
>
>
>-----Original Message-----
>From: David Fotland [mailto:fotland@xxxxxxxxxxxxxxxxx]
>Sent: Thursday, November 25, 1999 1:19 PM
>To: computer-go@xxxxxxxxxxxxxxxxx
>Subject: RE: computer-go: Applying Moore's Law to Computer Go
>
>how about this ?
>
>OOO
>OXO
>OAO
>XOX
>XXX
>
>The 'A' point capture one stone. but .....
>
>Simple ko is trival - move creates a new single stone group
>and captured exactly one stone.  Record the point of the capture
>as the "ko point", and the next ply ply moving there is illegal.  
>
>The rule should be:
>	the next time ply moving there alse capture one stone, it is
>illegal.
>
>During lookahead I only check for simple ko, not full board repetition.  
>
>David
>
>At 08:50 AM 11/23/99 MST, you wrote:
>>>My problem might be
>>>that, for every move, the program needs to do a lot of other stuff than
>>>simply evaluate the move. Activities include game control (who is next
>>>turn), rule validation (is it a legal move), board update (does it kill?
>>>update safeties of groups ...).
>>
>>I have this problem, too.  Especially calculating for "ko", because every 
>>call I make to IsValid() basically places the piece on the board, updates 
>>everything, and then checks to see if it is the same baord position-2.
>This 
>>takes way too much time.  Anyone found something better?
>>
>>Jeff
>>
>>______________________________________________________
>>Get Your Private, Free Email at http://www.hotmail.com
>>
>>
>
>