Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Pro Deo 1.0vff, Style by Eduard Nemeth

Author: Albert Silver

Date: 09:04:45 09/14/04

Go up one level in this thread


On September 14, 2004 at 11:50:33, Ed Schröder wrote:

>On September 14, 2004 at 10:17:40, Albert Silver wrote:
>
>>>That's because you are using the MISC_17 switch. It forces the engine to skip
>>>the first extension which generally will speed up the search between 20-50%.
>>
>>Any comments on the results of your testing of this?
>>
>>>One thing you should avoid, when using the MISC_58 switch realize that [Passed
>>>Pawns = xxx] is still active. In your style you use both:
>>>
>>>[Passed Pawns = 120]            * 120
>>>[Pruning = MISC_58]             * Recalculate passed pawns
>>>[M58_Table = strong]            * normal|midg|endg|strong
>>>
>>>This means that the STRONG setting is rewarded with an extra 20% which is way
>>>out of proportion. When using STRONG it's more likely that [Passed Pawns = 95]
>>>wil do a better job.
>>
>>I find that the STRONG is already out of proportion and use the NORMAL setting.
>
>I have the same feeling....
>
>
>
>>>For the next version I am planning a build-in help-text that explains all the
>>>parameters in detail, which textfile you can freely modify with your own
>>>comments.
>>
>>Yes! Please! :-)
>>
>>Could one pester you about three parameters all the same? (any answer would be
>>most welcome)
>
>>Is the Strong Squares only for knight outposts, or does it encourage advanced
>>outposts for pieces such as the bishop or rook as well?
>
>It's only for Knight outposts.
>
>
>>How does the Pawn Pressure parameter affect the game? Does it pawn advances or
>>is it something else altogether?
>
>[Pawn Pressure = xxx] measures the pressure on an opponent pawn that looks a
>target to attack, see for instance this (example) position:
>
>[d]rnbq1rk1/1p3ppp/p2p1n2/4p3/4P3/1NN5/PPP1BPPP/R2Q1RK1 w - -
>
>Pawn d6 looks an interesting target to attack, the algorithm will try to
>increase the pressure on square d6.
>
>The algorithm is split into 4 evaluation categories (game stages), the following
>setting is on my TO_TEST list.
>
>[Pawn Pressure midg = 150]
>[Pawn Pressure end0 = 200]
>[Pawn Pressure end1 = 200]
>[Pawn Pressure end2 = 100]
>
>
>>What is the difference between [Pawn Formation] and [Pawn Structure]?
>
>[Pawn Structure] is an overall parameter of the following parameters:
>
>[Pawn Pressure = 100]                * see explanation above
>[Traditional Isolated Pawns = 100]   * eval isolated pawns
>[Pawn Formation = 100]               * eval pawn formations
>[Weak Pawns = 100]                   * eval weak pawns
>
>So if you set [Pawn Structure = 120] it will set the above 4 parameters to 120.

Many thanks! As I like to tailor the above 4 parameters manually, I'll leave the
[Pawn Structure = 100] presuming that only the individual values will weigh in.
I suspect that I had the poor thing overdoing the Pawn business by adding
individual weights AND adding the Pawn Structure value to exacerbate the values
even more.

                                       Albert

>
>Hope you don't have a headache yet!
>
>My best,
>
>Ed



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.