Author: Uri Blass
Date: 09:05:46 04/24/04
Go up one level in this thread
On April 24, 2004 at 10:15:30, Djordje Vidanovic wrote: >On April 24, 2004 at 09:41:27, Uri Blass wrote: > >>On April 24, 2004 at 09:18:27, Djordje Vidanovic wrote: >> >>>On April 24, 2004 at 08:47:42, Kurt Utzinger wrote: >>> >>>>On April 24, 2004 at 08:07:48, Djordje Vidanovic wrote: >>>> >>>>>On April 24, 2004 at 07:32:32, Uri Blass wrote: >>>>> >>>>>>On April 24, 2004 at 07:04:44, Johan Havegheer wrote: >>>>>> >>>>>>>Hello >>>>>>> >>>>>>>Ruffian won against The King in a tremendous endgame >>>>>>> >>>>>>>To djordje : Everything is working fine, out off book against Hydra yesterday >>>>>>>with >>>>>>>+ 0.56 >>>>>>> >>>>>>>We lost against Hydra due to the depth and a wrong plan in the endgame >>>>>> >>>>>> >>>>>>What do you mean that you lost due to the depth. >>>>>>Did Ruffian need another ply to avoid the losing blunder? >>>>>> >>>>>>Uri >>>>> >>>>> >>>>> >>>>>Ruffian reaches ply 18 easily here, but it still clings on to gxh4. It needs >>>>>play 19 to see that things are not peachy. (BTW, Hydra was doing a 20 ply search >>>>>here.) But I doubt that a single CPU version of Ruffian could actually avoid >>>>>the move at the Leiden time controls anyway. As already pointed out some other >>>>>programs also go for the pawn. What does Movei do? >>>>> >>>>>Djordje >>>> >>>> I doubt that it's a matter of depth and would rather say >>>> it's a matter of chess knowledge. >>>> Kurt >>> >>> >>> >>>I tend to disagree with you here. I'm pretty sure that Ruffian has more endgame >>>knowledge than Hydra :-). Another issue crops up here -- and I think that we >>>will disagree on that one too: I think that deep search = knowledge (good >>>examples can be Deep Blue and, nowadays, Hydra). But this is more of a >>>philosophical issue and I'd like to spare both you and the CCC membership as the >>>issue can easily turn into a flame war. >>> >>>Thanks for the insight anyway. >> >>What about knowledge of better time management? >>gxh4 increase the number of passed pawns for both sides so you can decide to use >>more time when you plan to make a move that increase the number of passed pawns >>of both sides. >> >>Uri > > >You made a very good point. However, if I remember correctly, Ruffian used >quite a lot of time for gxh4 but could not find a better continuation. > >I also wanted to add that my comment regarding Hydra's endgame knowledge was >rather unfortunate and incorrect. It was based on speculation, as I presumed >that speed was what mattered most in Hydra. After talking to Alexander Kure on >playchess.com today I found out that Ch. Donninger had implemented lots of >endgame knowledge in Hydra, so that h4! was to be expected. Thus I apologise >for my unfounded speculation :-). > >Djordje Movei also cannot find something better even after hours on P850 inspite of seeing negative evaluation against white. depth=16 -0.16 g3h4 f5f4 b3c2 c7h7 c2d1 e5c6 c3h3 e6e5 d1e2 e5e4 h4h5 c6d4 e2f1 e4e3 h5h6 f6e6 f1e1 Nodes: 1579659624 NPS: 179573 Time: 02:26:36.75 Movei used only 32 mbytes hash but I do not think that it is very important and movei is not very sensitive for hash tables. Uri
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.