Author: James T. Walker
Date: 11:13:45 11/25/02
Go up one level in this thread
On November 25, 2002 at 10:38:14, Uri Blass wrote: >On November 25, 2002 at 10:25:35, James T. Walker wrote: > >>On November 25, 2002 at 06:22:49, Chessfun wrote: >> >>>On November 24, 2002 at 22:44:32, James T. Walker wrote: >>> >>>>On November 24, 2002 at 13:20:36, Chessfun wrote: >>>> >>>>>On November 24, 2002 at 11:09:48, James T. Walker wrote: >>>>> >>>>>>On November 23, 2002 at 13:46:33, Jan Kiwitter wrote: >>>>>> >>>>>>>On November 23, 2002 at 13:41:27, Jorge Pichard wrote: >>>>>>> >>>>>>>>On November 23, 2002 at 11:01:49, Christophe Theron wrote: >>>>>>>> >>>>>>>>>On November 23, 2002 at 11:00:04, Joachim Rang wrote: >>>>>>>>> >>>>>>>>>>On November 23, 2002 at 08:14:08, Jorge Pichard wrote: >>>>>>>>>> >>>>>>>>>>>Deep Fritz will have another miraculous return in the last 6 games just like in >>>>>>>>>>>the previous matches. I predict a 21 to 19 in favor of Chess Tiger 15. >>>>>>>>>>> >>>>>>>>>>>Pichard. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>where is the return, if you predict another 6 draws? >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>>:) >>>>>>>>> >>>>>>>>>Good point! :) >>>>>>>>> >>>>>>>>> Christophe >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>No, I didn't say that I predict 6 more draws, but probably another win or even >>>>>>>>two for Deep Fritz and the remaining draws. >>>>>>> >>>>>>>But then the result cannot be 21-19. >>>>>>> >>>>>>>> >>>>>>>> Pichard >>>>>> >>>>>>I'm amazed it's this close. I have played 25 games of Fritz 7 vs Chess Tiger 15 >>>>>>at G/60min and Fritz 7 leads by 18.5-6.5. Chess Tiger 15 is currently ranked # >>>>>>22 in my standard database after 101 games. I'm getting miserable results for >>>>>>Tiger 15-normal. I find it hard to believe it improves dramatically when the >>>>>>time control is raised to 40/2hours. >>>>>>Jim >>>>> >>>>>Going back to beta testing as you may recall, my results even at 40/40 were >>>>>always good, as were others. Also this match is Deep Fritz 7 not Fritz 7. >>>>> >>>>>I also don't see the time control as being too relevant as again from memory >>>>>even at fast blitz Tiger 15 performed better than Tiger 14. >>>>> >>>>>Sarah. >>>> >>>>Hello Sarah, >>>>Well if you go back to the beta testing you will see I never did have really >>>>good results. In fact vs Junior 7 I seem to remember that Tiger got like 3.5 >>>>points out of 24 games (7 draws no wins). It was close to Fritz 7 but never was >>>>ahead of Fritz 7 in my test. I played over 1000 games and my results were >>>>mediocre especially vs Fritz 7 & Junior. As I said above Chess Tiger 15 was >>>>ranked # 22 in my database. That's not very impressive although it contains >>>>only 101 games of Chess Tiger 15-normal. Since the beta test I have now >>>>swithced to XP2400+ cpus and so far results are about the same. I'm running >>>>Tiger with either 96M/192M hash and have all 3/4/5 man tablebases. It also beat >>>>Ruffian by only 14-12. Ruffian is ranked about #8 in my database now and Chess >>>>Tiger 15 moved up to the 18th spot after some wins vs Hiarcs 8 today. I don't >>>>like to test at 40/2hrs because it takes too long and besides the SSDF does a >>>>good job at that time control. >>>>Jim >>> >>>I remember your results against Junior 7 as it was after the final release and >>>as I recall we had a thread on the subject. As you remember I also tried exactly >>>at the same time as you and posted my 9 games at 40/40 +3 -2 =2. >>> >>>If I also remember your results earlier were much better although there were no >>>changes in the final beta. I even took the liberty of checking and your results >>>against Junior 7 in longish games were; >>> >>>Chess Tiger 15 normal vs Junior 7 >>>@G/60 min 9.5-4.5 >>>@ 40/60 min 7.0-3.0 >>> >>>Taken from your post dated 08-09-2002. >>> >>>Sarah. >> >>As I said, I played over 1000 games. That post was after 925 games and ignores >>the poor results with Junior mentioned earlier. My results since then have >>continued to be worse as indicated by my post. My latest database does not >>include the beta test games. I never included them in the database because it >>was a "beta" version as far as I knew and I didn't want to screw up my database >>with games from a broken beta engine. In retrospect I could have kept them >>since the beta ended up being the final version as far as I know. Here is what >>is in my "standard" database now for Tiger 15 vs: >>Fritz 7 8.5-16.5 (corrected) >>Junior 7 2.0-12.0 >>Ruffian 14.0-12.0 >>Shredder 6Classic 7.0-11.0 >>Hiarcs 8 9.0-9.0 >>Shredder Paderborn 5.0-3.0 >>Aristarch 4.4 4.5-2.5 >>Junior 6 3.5-1.5 >> >>That is a score for Tiger 15 of 53.5 out of 121. Not very impressive. The >>above results vs Fritz 7 and Junior 7 are terrible. I don't know if it's the >>result of "learning" by those two during the beta period or what. I doubt it's >>that. It could just be bad luck but it started way back near the end of the >>beta period. I have no explanation for the results, only the results >>themselves. > > >The first explanation to think about is that it is possible that tiger's >computer was slower for some reason. > >I think that it may be productive if you have the average number of nodes per >second for tiger and other programs. > >Uri Hello Uri, I'm running 2 XP2400+ computers. Here is a sample of their performance using the position posted by Terry M. above. The times indicate they are within <2% of each other and this is while accessing the HD for tablebases. Computer "A" 00:04:18.8 0.00 17 151510584 Qa6+ Kb3 Qb5+ Ka2 Qa4+ Kb1 Qd1+ Ka2 Qa4+ Kb1 Qd1+ Ka2 Qa4+ Kb1 Qd1+ Ka2 Qa4+ Kb1 Computer "B" 00:04:15.3 0.00 17 151510584 Qa6+ Kb3 Qb5+ Ka2 Qa4+ Kb1 Qd1+ Ka2 Qa4+ Kb1 Qd1+ Ka2 Qa4+ Kb1 Qd1+ Ka2 Qa4+ Kb1
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.