Author: Robert Hyatt
Date: 17:58:22 10/17/00
Go up one level in this thread
On October 17, 2000 at 13:04:33, Thorsten Czub wrote: >cannot find the original thread anymore, it scrolled away due to >the kramnik-kasparov "noise"... > >again: what is wrong with this data ?Chess Tal 2.03 gegen die Anderen > >Antworten * Antwort schreiben * Computerschach-Forum > >Geschrieben von Rudolf Klamert am 14. September 2000 um 13:02: >Chess Tal 2.03 gegen die Anderen >Hardware: >2 x AMD-K6/III-450 je 256 MB SDRAM >Motherboard: 2 x EP-MVP3G 1MB Cache >Bücher, Tablebases und Talk auf Platte. >Je 10 Autoplayer-Partie 40/120 + 20/60 Rest 30 >Gespielt werden nur Einzelpartien mit wechselnder Farbverteilung. >Beide Rechner werden nach jeder Partie neu gebootet um evtl. Fehler >durch TBS im Cache ( Speicherfreigabe W95/98 ! ) zu vermeiden. > >Chess Tal 2.03 mit opening.bok + large.book. Einstellung Strength >Hashtables: Main Hash 131.072 kB, Pawn Hash 65.536 kB >Crafty 17.10 mit Powerbook 2000 Hashtables 128 MB >Fritz 6.0a mit PowerBook 2000, Hashtables 184 MB >Fritz 5.32 mit PowerBook 2000, Hashtables 184 MB >Fritz 5.00 mit PowerBook 2000, Hashtables 131072 kB >Junior 6.0 mit jbook.ctg, Hashtables 184 MB >Nimzo 7.32 mit nz7book.ctg, Hashtables 184320 kB >Nimzo`99 mit nz7book.ctg, Hashtables 18448 kB >Hiarcs 7.32 mit Bookupdate ( General.ctg + Hiarcsbook ) Hashtables 188416 kB >Shredder 4.0 mit shredder.bk7 Hashtabellen 96 MB, Bewertungtabellen 32 MB >Genius 6.5 mit Mainbook.gbd Hashtables 128 MB >Century mit Rebel.mvs und 60 MB Hashtables >Century 1.2 mit Rebel.mvs und 60 MB Hashtables. Reiner DOS-Mod >Hiarcs 7.01 Eröffnung Turnier. Hashtables 65472 kB. Reiner DOS-Mod >Ergebnis: > >Chess Tal 2.03 - Crafty 17.10 ---------------- 4 zu 3 und 3 x Remis = 5.5 zu Several things are wrong with this (and all remaining) pieces of data here: 1. 10 games is too small to draw _any_ conclusion from. 100 would give me much more confidence in the results. 2. CSTal supposedly represents a "new paradigm" yet it beats me by 1 point in a very short match? That isn't convincing me of anything at all, other than that the match was too short. If after 100 games you are still 1 game ahead, the "new paradigm" isn't giving you very much when we compare lines of code, evaluation, and the like??? 3. I've never seen CSTal attack me very well. Even when Chris and I played some games with my king safety 100% turned off. I was essentially fast enough to avoid most problems _anyway_. If programs that are even faster are used (tiger comes to mind) how would CSTal fare there? I'm not saying what he is doing is garbage at all, and didn't mean to imply that if I did. I am simply saying that the "less speculative" approach also has its advantages... I would rather enter an endgame with an equal position, as opposed to entering the endgame a piece down after a speculative attack failed. In the latter case I am lost, in the former I still have a chance. 4.5 >Chess Tal 2.03 - Fritz 6.0a -------------------- 4 zu 5 und 1 x Remis = 4.5 zu >5.5 >Chess Tal 2.03 - Fritz 5.32 -------------------- 2 zu 3 und 5 x Remis = 4.5 zu >5.5 >Chess Tal 2.03 - Fritz 5.00 -------------------- 2 zu 3 und 5 x Remis = 4.5 zu >5.5 >Chess Tal 2.03 - Junior 6.0 -------------------- 2 zu 3 und 5 x Remis = 4.5 zu >5.5 >Chess Tal 2.03 - Nimzo 7.32 ------------------ 3 zu 5 und 2 x Remis = 4.0 zu 6.0 >Chess Tal 2.03 - Nimzo`99 -------------------- 2 zu 3 und 5 x Remis = 4.5 zu 5.5 >Chess Tal 2.03 - Hiarcs 7.32 ------------------ 2 zu 5 und 3 x Remis = 3.5 zu >6.5 >Chess Tal 2.03 - Hiarcs 7.01 ------------------ 2 zu 7 und 1 x Remis = 2.5 zu >7.5 ** >Chess Tal 2.03 - Shredder 4.0 ---------------- 1 zu 6 und 3 x Remis = 2.5 zu 7.5 >Chess Tal 2.03 - Genus 6.5 ------------------- 2 zu 5 und 3 x Remis = 3.5 zu 6.5 >Chess Tal 2.03 - Rebel Century --------------- 7 zu 1 und 2 x Remis = 8.0 zu 2.0 >Chess Tal 2.03 - Rebel Century 1.2 ----------- 1 zu 7 und 2 x Remis = 2.0 zu 8.0 > >1 Hiarcs 7.01 ------------- 2626 ------------- 1 0 1 1 ½ 0 1 1 1 1 -------- >7.5/10 >2 Chess Tal 2.03,A ------- 2529 ------------- 0 1 0 0 ½ 1 0 0 0 0 -------- >2.5/10 >Wird fortgesetzt. >Grüße aus München >Rudolf Klamert
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.