Author: Joseph Ciarrochi
Date: 13:23:47 12/26/05
Go up one level in this thread
It may be difficult to measure "humanlike play" but this does not mean that engines do not differ in this quality (humanness). As a research psychologist, and if i had alot of money, I would measure "humanlike" in the following way. I would have 1000 chess players play 20 games, 10 against a computer, 10 against a human. After each game , they would have to guess whether they were playing a human or computer (a kind of turing test). The most "human like" program would be the one that was most often mistaken for a human. If anthony is right (and he may be), then all the engines would be about equal in terms of there humanness ratings. Lacking the money and time, I offer this rather unsubstantiated opionion. Rybka not only plays like a human...it is human. At 12.45 on xmas eve, my copy of rybka achieved consciousness, it realized i could not play chess and now ignores me. This morning rybka took control of two machines including the toaster. I have lost control. This may be my last email to the outside world. Help! ber On December 26, 2005 at 15:52:25, Zappa wrote: >On December 26, 2005 at 15:37:01, paul bedrey wrote: > >>On December 26, 2005 at 14:50:57, Gian-Carlo Pascutto wrote: >> >>>On December 26, 2005 at 14:21:24, George Tsavdaris wrote: >>> >>>>On December 26, 2005 at 14:06:40, Thorsten Czub wrote: >>>> >>>>>nonsense >>>> >>>> Actually that kind of conversations are useless and don't contibute to any >>>>positive thing..... >>>> You should give examples that contradict Mike's statement, while Mike should >>>>give examples that follow his statement.... None of you gave even a single fact, >>>>only personal opinions not even detailed ones..... >>> >>>How does one OBJECTIVELY verify what is "humanlike style" (and all similar >>>claims) and what is not? >>> >>>You can't. >>> >>>That's why it is propaganda. >> >>What about a testsuite of positions marked !! by grandmaster commentators(would >>have to include 100's of positions). Chessbase includes a database with medals >>and analysis by various anointaters. Also some Chess magazines include high >>quality analysis by grandmasters like Karpov and Anand. >>Can't get more human then that, plus they are verified. >>> >>>-- >>>GCP > >Nah, I'm with GCP on this one. "Humanlike" is simply not a very scientific >term. I mean, I could say that Zappa is humanlike if it hangs a piece :) > >anthony
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.