Author: Vincent Diepeveen
Date: 09:27:01 04/14/02
Go up one level in this thread
On April 14, 2002 at 12:18:40, Uri Blass wrote: what i have against nunn's choice is at the time he chose them it was favouring chessbase fritz engine obviously (tiger wasn't sold by chessbase back then, note that i never got tiger to work at analysis mode under chessbase GUIs). it was presented back then as positions to play out, giving impression they were not book positions, where in fact they were bigtime inside the book. another thing is the type of the positions. Please look carefully what kind of positions they are and compare with what type of positions you have in all the different games you get on the board in your corr games. Exactly. It is a certain *type* of positions chosen by nunn where innocent shuffling in combination with book is simply scoring points. The positions are *that* complex that knowledge doesn't necessarily work in the positions, because basic chess knowledge which for example Tiger has but not Fritz, that knowledge simply isn't rewarding in some of the positions. Knowing even less than less is rewarding in some of the positions, because the strategic idea is simply too hard for *any* engine. In short the positions were completely not representative. They represent something which engines do not understand, so that means in short that if an engine X knows a bit more than other engines, that engine X is not having an advantage here. In that respect the positions WERE carefully chosen. Let's still not discuss the possible aspect of commercial advantage. If my company would pick a chessset to test upon and if i then am allowed to present as the first one the results, then i of course first will have fixed some things. Like some lines in my book for that position. In that last aspect the book is still dominating in the positions. >On April 14, 2002 at 11:29:39, Vincent Diepeveen wrote: > >>On April 13, 2002 at 16:23:12, Kurt Utzinger wrote: >> >>congratulations Christophe, chessbase beaten in its >>own tournament! I remember that at the time the positions >>were chosen it completely favoured fritz. All positions >>were closed shuffle positions. At that time fritz was not >>worse there than other programs. > >Fritz is still leading the tournament. >Tiger won only one position. > >Tiger is also a chessbase engine >The king is the only engine in the tournament that is not sold by chessbase. > >> >>Also i was a bit sick that he used positions which still were >>completely INSIDE the book. > >I see no problem with it. >The tournament only shows which programs are better >in some specific positions. > >> >>Nunn obviously knows nothing from computerchess here, already >>years ago we figured that out. > >I do not understand what do you have against nunn. >It is clear that we cannot say which engine is better based on 10 positions >and it is not nunn's fault. > >It is possible to optimize engines for every known position so if people want to >test engines without opening book in order to see which engine is better in >analyzing they should always change the positions(it is possible to test engines >in the positions after 15 moves of smirin in the next 8 games of smirin against >the machines and for the next year it is possible to use different games in >order to prevent preperation) > >Uri
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.