Author: Dieter Buerssner
Date: 10:52:51 02/01/04
Go up one level in this thread
On February 01, 2004 at 12:41:53, Robert Hyatt wrote:
>On February 01, 2004 at 07:20:20, Dieter Buerssner wrote:
>
>>On January 31, 2004 at 22:08:33, Robert Hyatt wrote:
>>
>>>On January 31, 2004 at 21:37:29, J. Wesley Cleveland wrote:
>>
>>>>>What do you do with KRPP vs KR positions deep in the tree where you need to
^^^^^^^^^^^^^^^^
>>>>>promote?
>>>>
>>>>I take the exact value from the KRPP vs KR table and prune. You only need
>>>>the KR** vs KR tables if you promote with a capture.
>>>
>>>How can you do that? How do you know the promoted position is a win or a draw
>>>(or even a loss) even though the previous un-promoted position was a win or
>>>draw?
>>
>>You don't need to know it inside the search tree. As Wesley said, you have
>>already cut the tree, before you are going to promote.
>I don't follow. I reach a krppkr position at the root, and I make the best
>move, until there is no winning move except a promotion which I can't see. So
>now I do a search, and I _still_ need the tables to make sure I don't promote
>wrongly into a draw...
>
>If you have krpp vs kr, you still reach positions where you will have to search
>or else shuffle and draw. And there you need the right tables or you can end up
>with =Q and a draw where =N may win...
But you won't need it inside the search. I thought this was, what you were
discussing (see my ^^^^^^) above. When the position is at the root, things will
of course be different. Like Uri, I would think that in almost all cases a
normal search (don't probe KRPPKR now, but you still can probe all 5-men TBs and
other 6 men TBs) should find the move always almost.
Regards,
Dieter
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.