Author: Jeremiah Penery
Date: 13:42:40 10/16/99
Go up one level in this thread
On October 16, 1999 at 10:24:23, James T. Walker wrote: >On October 15, 1999 at 09:53:51, Jeremiah Penery wrote: > >>On October 15, 1999 at 03:44:30, Ed Schröder wrote: >> >>>>Posted by Marcus Kaestner on October 14, 1999 at 04:01:40: >>>> >>>>in chessbits 3 i´ve made a test with the nunn-endgames (10 endgame-positions >>>>played with both sides). every program one time with, one time without tb´s >>>> >>>>after abaout 140 games >>>>h7.32 earned +5 elo for the tb-version >>>>n7.32 earned +12 elo >>>> >>>>far the best was crafty 16.13 with +65 elo, topping even the score of hiarcs >>>>7.32tb >>>> >>>>marcus >>> >>>Take this into account: Some programs may have so much chess knowledge in their >>>programs that (the current) TB's have zero effect in the end. This will change >>>as soon as the 6-man TB's (and up) will come, no doubt.. >> >>Definitely. >> >>>Other programs are developed to 100% rely on TB's and therefore have very few >>>(or nothing) chess knowledge inside because there is no need for that. I don't >>>know in to what extend this is true for Crafty but I am confident Bob will >>>tell you. >> >>I don't think there are any programs developed to "100% rely on TB's" so that >>they wouldn't have any knowledge about the endgame. Whatever search/eval they >>have will still be quite good (At least in endings where TBs would affect things >>- 5 piece or less), it's just that the TBs provide 100% accuracy. >> >>>As a result you can not compare the value (elo gain) of TB's in this way as >>>it will depend of what is already inside a program (or not). >>> >>>Rebel knows the most important ones such as (by head): >>> >>>- KRK >> >>Yes, this one is easy. >> >>>- KNNK >> >>Of course, this is always a draw. >> >>>- KBNK >> >>This one's interesting, but not really that difficult. >> >>>- KPK >> >>This is pretty easy, too. >> >>>- KRPKR >>>- KNPKN >>>- KBPKB >>>- KQPKQ >> >><lumping these last 4 together> >>I'm a bit skeptical about some of these. I'm sure it may play even 90% >>accurately in these endings, but I'd be willing to bet it couldn't win 100% of >>winnable positions from these endings (especially KRPKR and KQPKQ) against an >>opponent with the TBs. >> >>>I am sure this list is not complete. Of course the bad-bishop situation is >>>covered too. Some others are not covered such as KQKR (Hi Jim!) for which >>>Rebel will rely on TB's in the future. >> >>I'm sure Rebel could easily win KQKR without TBs, even against an opponent with >>them. It's an easy ending for computers. >> >>Jeremiah > >Hello Jeremiah, >I wish this were true! See the (Hi Jim) above. I brought this up to Ed last >week. I saw Bob's post concerning this ending as trivial since Crafty could do >it easily. I decided to test this and gave Crafty a Mate in 28 against Hiarcs >7.32. Crafty with no tablebases mated in 36 moves against Hiarcs with >tablebases. When I gave this to Rebel it failed badly. In fact at the 50th >move according to Hiarcs it was at "Mate in 28". Right where it started even >though at one point it had it down to about mate in 17. This is about the same >area where humans were found to have problems with this ending. It seems if >they could get past the mate in 15/16 area they could complete the mate. >Problem is most could not break this "Barrier". >Jim Walker Oh, my. I wonder why Rebel failed this so badly. Since it seems to be quite trivial for Crafty, but not so for Rebel, I wonder how other programs do at this ending. Thanks for the information. :) Jeremiah
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.