Author: Robert Hyatt
Date: 11:07:45 04/01/03
Go up one level in this thread
On April 01, 2003 at 04:59:42, Uri Blass wrote: >On March 31, 2003 at 15:38:51, Robert Hyatt wrote: > >>On March 30, 2003 at 05:05:34, George Sobala wrote: >> >>>On March 29, 2003 at 16:52:21, Jan Kiwitter wrote: >>> >>>>On March 29, 2003 at 15:17:01, Jason Waugh wrote: >>>> >>>>>It's style is highly customizable - it comes with a bunch of different >>>>>personalities pre-configured, along with a personality editor to make your own. >>>>>RE: strength.... I wouldn't buy it for it's strength compared to engine 'X' - >>>>>but it makes an excellent complement to Chess Tiger 15 because it is so much fun >>>>>to play --- DS for playing, CT15 for analysis. >>>>> >>>>> >>>>>Jason. >>>> >>>>Thanks for you opinion! Why is it fun to play with? Does it play very active? >>>>Regards >>>>Jan >>> >>>You can make it play VERY active! See redshift running on ICC for an example - >>>playing to a rating of about 2450 on a K6-450, very happy to make sacs in the >>>opening and middle game. Other engines slaughter it because they defend >>>precisely, but against humans it is different: e.g. against one IM (Chito Garma, >>>FIDE 2415) who has played it at 3 0 blitz on this setting it has scored +61 =3 >>>-46. The very fact that he has played it 110 times suggests that there must be a >>>fun element! >> >> >>yes, but that -46 suggests that the settings while "fun" are very "bad". > >Everything is relative > >+61 suggests that the setting was better than the 2415 at 3 0 game. > >I think that some change in the rules is needed if you want to do blitz games >of GM's against computers interesting and fair. > >A possible idea that I suggested in the past is to allow both sides to take back >moves and times limited number of times in the game. > >I think that supporting this idea by chess servers is not hard to do. >I do not know how many take backs are needed for GM's to score 50% against >computers in blitz. > >Note that take back could be also without playing a move and if a player thinks >a long time about interesting option and suddenly see that he has almost no time >left he can ask to take back the time instead of playing a wrong move and asking >to take back move and time. > >Uri For a computer to almost break even at 3 0 is _terrible_. I'd play any human on the planet at 3 0 and expect to win at _least_ 9 out of every 10 games, and I have watched GM players lose 100 in a row at 3 0. Yes, it might draw one here and there, and it might lose one every 20 or 30 games at worst. But game in 3 minutes is so favorable to the computer as to be considered almost unplayable by anyone in the world. There is no take-back using xboard/winboard/etc on ICC, so that doesn't happen. But I'd be willing to bet that to win at 3 0, they are going to need _several_ takebacks in a single game, and no GM player even asks for takebacks on ICC...
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.