Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: levers and pawnhash hit-rate

Author: Gerd Isenberg

Date: 13:50:03 11/17/04

Go up one level in this thread


On November 16, 2004 at 11:36:44, Robert Hyatt wrote:

>On November 15, 2004 at 15:57:23, Gerd Isenberg wrote:
>
>>Hi,
>>
>>i use the difference of weighted pawnbitboards ( 5*whitepawnBB - 3*blackpawnBB )
>>mod 65077 as an index for a 64K entry pawnhash-table. Most often my pawnhash
>>hit-rate is quite ok, > 95% or even 99%. But it seems that a lot of pawn levers
>>decrease my hit-rate due to collisions.
>>
>>With "maximum lever" positions my pawnhash hit-rate is only about 75% after a 30
>>second search on my amd64 2.2GHz box (~80% without eval hashing).
>>
>>I guess a zobrist-key pawn-hashing scheme is better here.
>>Curious about your pawnhash hit-rates with one of the positions below.
>>
>>Btw. how many levers has each position, 8 or 14?
>>
>>Thanks,
>>Gerd
>>
>
>For the first position, I don't consider it to have _any_ levers.  I evaluate
>(personally) levers to be pawn pushes that cause contact with an enemy pawn.
>These pawns are all in contact already.  The idea behind a lever is that it is a
>_potential_ pawn break that you can play when you are ready...

I see - AFAIK Kmoch's definition of a pawn lever (hebel) is a pair of pawns in
capture distance with the notation eg. e4:d5 or from black point of view d5:e4.
Similar rams (widder) with Kmoch's notation e4/e5.

May be a matter of translation, because i refer on Hans Kmoch's
german terms in "Die Kunst der Bauernführung".

I agree that your definition, the ability to build a lever by a pawn push, a
pawn break, is (more?) worth to consider in eval as well.

IMHO with Kmoch's definition, there are 14 levers and 8 rams (as a pair of
white/black pawns) in those posted positions. Each rammed pawn is also member of
the set of lever- or even double-lever- (left/right) pawns, which makes the
rammed property, which is usually a sign of congealment, unimportant.

Pawn levers or pawn-breaks are usually more worth, if the opposite "lever"-pawn
is already rammed (or blocked) and has no other option to capture or to get
captured. But of course one has to distinguish backward pawns, candidates and
whether (half) opening this file is advantageous.

"Eternal" rammed pawns without the ability of opponent pawns to build a lever
are of course good for "defence", e.g. as part of a king shelter.

Gerd

>
>>
>>[D]8/k7/8/pppppppp/PPPPPPPP/8/8/K7 w - -



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.