Author: Vincent Diepeveen
Date: 14:40:39 04/04/01
Go up one level in this thread
On April 04, 2001 at 15:25:01, Ulrich Tuerke wrote: >Hi Vincent, > >yes, in some way the gambit tiger seems to polarize the feelings of the users. >If they are not bloodily annoyed, then they are incredibly happy. Right, and if you release a free new version of Comet or whatever i bet you hardly get the same reactions. >The programmers know that even Chris is using just water to boil his coffee. >Nevertheless, it's certainly an interesting program and I'm looking forward >trying it. But I doubt that it's so totally different than the others. >I am sure that even Diep has some 2nd order evaluation terms, right ? Actually i'm pretty sure i invented that term some years ago when discussing mobility and why i wasn't happy by just a summation of the squares that were occupied by a piece, as i wanted to do piece <-> piece relationships (also piece <-> pawns) and piece <-> attacks relationships evaluated. However i was later corrected by Frans Morsch (who prefers to not post here because he otherwise gets a zillion happy+negative reactions more as Christophe nowadays is getting here) that in fact the mobility in itself was already second order and that the rest was even higher order evaluation. So we can also discuss in length what is second order is exactly, as very important for chess IMHO are scanning patterns, like mobility is, whereas a relationship like BISHOP_D3_WEAK is a pretty simple pattern, though it depends upon a pawn on d2. So there is already a relationship there. Yet last period of time my conditions to apply a pattern/scan has only increased for the average pattern, so i guess it's quite smart to say that i'm above second term already a long period of time. A difficult term also is of course pawn majoritycode. As i need 3 loops inside each other to figure it out and it has many other piece relationships of course. Sometimes a majority is everything but good... What order would 3 loops in each other be? Anyway let's not talk too much about other complex patterns i have too much, as despite much comments i get. I mention 'pawn majority' once and DANG within a year it is also mentionned by Ernst A. Heinz (if i remember well) in a Deep Blue article and it seems that all terms i mention here get used some time later. Can't blame anyone however. Chess players suck incredible in naming what they see. Like in chess there is the term 'good bishop' and 'bad bishop'. But you can have a good bishop, but also an excellent bishop. Like in dutch i call sometimes a bishop a 'positions bishop'. Like the c8 bishop in KID is. If you lose it, then you're dicked forever usual... Evaluation in general is a complete unexplored territory for scientists and chess players. In some books people have done tries and efforts to name things, but they are very bad in naming positional factors, whereas there are plenty of names for tactical things (probably because humans make so many mistakes with it), which for a part are completely useless as search will pick it up anyway. >Greetings, Uli
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.