Author: Eelco de Groot
Date: 18:18:53 10/29/00
Go up one level in this thread
On October 28, 2000 at 18:35:32, Uri Blass wrote: >On October 28, 2000 at 17:59:23, Timothy J. Frohlick wrote: > >>GT did consider Bb7 for about one minute after 7 minutes of thinking on my PII >>333 with 28 Mb hash. The analysis was similar to Sarahs' afterward. The score >>did remain negative for black however. >> >>GT would never "under-develope" to the extent that black did. The game is a >>draw even with material superiority due to the horrible back rank situation. At >>least GT tells us how "not" to play. >> >> >>Tim Frohlick > >I posted the game. >I do not know if gambit would get itself to the position that I posted(I think >that bxc5 was too greedy and Ba6 was better but I did not analyze the game for a >long time and I may be wrong). > >You can try to see if there is a move in the game that gambit does not like. > >I see that gambit understands the position better than other programs(it can at >least see that f6 has negative score) > >It cannot see that Bb7 is good enough for black to draw but I am not sure if Bb7 >is really good enough for a draw(I only know that it was probably good enough >for moshe to draw and it is possible that gambit can improve the opponent's >moves after Bb7) > >Uri Hi Uri, Maybe Tim is right and the game should be a draw. What I would like to see is a good winning line for White after 14. .. f7-f6 or 14. ..f7-f5, that could tell us more. Rebel Century 2.0 favors White after 14. ..Bc8-b7 but not after 14. ..f6 after a not very long analysis. I'm not sure if a positive or negative score really means that much. Q5T thinks White is better after 14. ..f6 which is what you would maybe want to see, but that is maybe only because it values any attacking chances in the position higher. It only makes a difference for a program if because of the different evaluation it chooses different moves and it somehow can thus discover a winning line for White. But in this case they still play much the same moves so a positive or negative evaluation does not say enough, if there isn't a probably decisive line going with it as well. Which Q5T at least hasn't seen yet after 14. ..f6 or the score would go down further. On the other hand, about another post below about "game #63", I'm also not convinced that an aggressive program like Roy's 'Storm' could not be good at defending too. The scores might favor the attacking party but if there are good defending moves it would probably find them. Maybe a little slower than an allround program if you first prefed it the attacking moves, but I doubt that. Maybe there could be a problem if somehow the defending party has the possibility of counterattacking and it could choose to do that too soon, just as it could be too fast with an attack in other situations. I find that Q5T normally is able to refute its own stupid attacks given enough time. Because of the DEEPER search it reaches the higher plies needed for that, -it is more selective than default so that can be a reason it does not see some things at shallower plies-, not very quickly if it needs twelve ply for instance and has not much hash memory available. Storm does not use DEEPER search and also has default Selectivity so that is very different again. I think it is worth investigating how good these personalities and attacking programs are at also finding good defences, regardless of possible scores favouring the attacker. They go look for attacking chances for both sides which should mean that they can find the sharp lines in a position quicker. And sometimes they overvalue them or they miss something vitally important there. Gambit Tiger is probably much better than Q5T about not getting itself in a fix by too optimistic attacks. Rebel Century 2.0 after 14. ..f7-f6: Last Move : 1... f6 (White to play) 00:02 06.01 0.27 1..Nc6 2.Bf6 Nxf6 3.exf6 e5 4.fxg7 exd4+ 5.Kf1 Kxg7 00:00 04.01 -1.60 1.Bh6 Rf7 3.Bxg7 Rxg7 4.exf6 Nxf6 5.dxc5 00:00 05.00 -1.74 1.Bh6 Rf7 3.Bxg7 Rxg7 4.exf6 Nxf6 5.dxc5 Ne4 00:02 06.00 -1.16 1.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.exf6 Nxf6 6.Qh8+ Ng8 00:05 07.00 -1.16 1.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.exf6 Nxf6 6.Qh8+ Ng8 7.dxc5 00:16 08.00 -1.16 1.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.exf6 Nxf6 6.Qh8+ Ng8 7.dxc5 01:31 09.00 -0.48 1.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf7 5.exf6 Nxf6 6.Ne5+ Kf8 7.Qh8+ Ng8 01:57 10.00 -0.94 1.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.Qh8+ Rg8 6.exf6 03:32 11.00 -0.48 1.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf7 5.exf6 Nxf6 6.Ne5+ Kf8 7.Qh8+ 09:37 12.00 -0.45 1.Bh6 Qe7 3.Bxg7 Kxg7 4.Qh6+ Kf7 5.Nh4+ Rg8 6.Qh7+ 28:33 13.00 -0.45 1.Bh6 Qe7 3.Bxg7 Kxg7 4.Qh6+ Kf7 5.Nh4+ Rg8 6.Qh7+ 21:23 14.00 -0.48 1.Bh6 Qe7 3.Bxg7 Kxg7 4.Qh6+ Kf7 5.exf6+ Qxf6 6.Ng5+ Ke8 7.Nxe6 Qxf2+ After 14. ..Bc8-b7: Last Move : 1... Bb7 (White to play) 02:20 11.00 2.31 4..cxb2 5.Rb1 Qa5+ 6.Kd1 Nbd7 7.Qe6+ 00:00 03.00 -2.64 1.Bh6 Bxf3 3.Bxg7 Bxg2 00:00 04.00 -0.47 1.Bh6 Bxf3 3.Bxg7 Bh5 4.Bxf8 Qxf8 00:01 05.10 -1.00 1.Bf6 Nxf6 3.exf6 Bxf3 4.fxg7 Kxg7 5.gxf3 cxd4 6.Qh6+ 00:03 06.00 0.34 1.Bf6 Nxf6 3.exf6 e5 4.fxg7 exd4+ 5.Be2 Kxg7 6.Qxc5 00:06 07.00 0.46 1.Bf6 Nxf6 3.exf6 Bxf3 4.fxg7 Kxg7 5.Qh6+ Kf6 6.Qf4+ Ke7 7.Qxf3 00:13 08.00 0.30 1.Bf6 Nxf6 3.exf6 Bxf3 4.fxg7 Kxg7 5.Qh6+ Kf6 6.Qf4+ Ke7 7.Qxf3 Nd7 00:29 09.00 0.41 1.Bf6 Nxf6 3.exf6 Bxf3 4.fxg7 Kxg7 5.Qh6+ Kf6 6.Qf4+ Ke7 7.Qxf3 Qc6 8.d5 01:33 10.00 0.70 1.Bf6 Nxf6 3.exf6 Bxf3 4.fxg7 Kxg7 5.Qh6+ Kf6 6.Qf4+ Ke7 7.Qxf3 Nc6 02:55 11.00 0.34 1.Bf6 Nxf6 3.exf6 Bxf3 4.fxg7 Kxg7 5.Qh6+ Kf6 6.gxf3 cxd4 7.Qh4+ Kg7 8.Qxd4+ e5 9.Qe4 09:27 12.00 0.54 1.Bf6 Nxf6 3.exf6 Nd7 4.fxg7 Kxg7 5.Qh6+ Kf6 6.Qf4+ Kg7 Q5T Experimental after 14.Qc1: 00:00 05.00 1.49 1..cxd4 2.Qf4 f6 3.exf6 Bxf6 4.Nxd4 Nc6 00:01 06.00 -1.01 1..cxd4 2.Bf6 Nxf6 3.exf6 Qa4 4.fxg7 Qb4+ 5.Kf1 00:07 06.01 0.68 1..f6 2.exf6 Nxf6 3.Bh6 Qe7 4.Qg5 cxd4 5.Qxg6 Qf7 00:33 07.00 0.00 1..f6 2.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.Qh8+ Rg8 6.Qh6+ Rg7 7.Qh8+ Rg8 8.Qh6+ 00:37 07.13 0.37 1..f5 2.Qf4 cxd4 3.Qh4 Rf7 4.Qh7+ Kf8 5.Nxd4 00:40 08.00 -1.32 1..f5 2.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.Ng5 Nb6 6.Qh8+ Rg8 00:49 08.01 0.00 1..f6 2.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.Qh8+ Rg8 6.Qh6+ Rg7 7.Qh8+ Rg8 8.Qh6+ 01:19 09.00 0.00 1..f6 2.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.Qh8+ Rg8 6.Qh6+ Rg7 7.Qh8+ Rg8 8.Qh6+ 02:34 10.00 -0.82 1..f6 2.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.Qh8+ Kf7 6.exf6 Qxh8 7.Rxh8 Kxf6 06:49 11.00 -0.82 1..f6 2.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.Qh8+ Kf7 6.exf6 16:41 12.00 -0.74 1..f6 2.Bh6 Rf7 3.Bxg7 Rxg7 4.Qh6 Kf8 5.Qh8+ Kf7 59:17 13.00 -0.96 1..f6 2.Bh6 Qe7 3.Bxg7 Kxg7 4.Qh6+ Kf7 Sorry for this lengthy post here guys. Please, cheer up, a little! Kasparov will make a come-back, real soon, I'm positive about it, he has done it before. Regards, Eelco
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.