Author: Uri Blass
Date: 14:49:59 09/15/05
Go up one level in this thread
On September 15, 2005 at 17:07:11, Uri Blass wrote: >On September 15, 2005 at 17:03:28, Andreas Guettinger wrote: > >>On September 15, 2005 at 14:25:17, Ed Schröder wrote: >> >>>On September 15, 2005 at 14:07:05, Robert Hyatt wrote: >>> >>>>On September 15, 2005 at 07:21:19, Ed Schröder wrote: >>>> >>>>>On September 15, 2005 at 06:43:22, Gian-Carlo Pascutto wrote: >>>>> >>>>>>On September 15, 2005 at 06:26:30, Ed Schröder wrote: >>>>>> >>>>>>>On September 15, 2005 at 04:30:38, Mark Young wrote: >>>>>>> >>>>>>>>On September 14, 2005 at 17:05:59, Ernst Walet wrote: >>>>>>>> >>>>>>>>>Because Zappa 1.1 is SMP and 64 bit, I gave it a try and was pleasant suprised. >>>>>>>>>One of my favourite chess positions is this one, where Kasparov played Rxd4. >>>>>>>>>How do other engines do? If your engine finds it please give it a couple of more >>>>>>>>>plies to see if it holds it. >>>>>>>> >>>>>>>>The better test for the engine in this position, is to see when the program see >>>>>>>>that Rxd4 is not winning. Since Rxd4 is only a draw with best play. >>>>>>>>This is why many programs will not play Rxd4, or drop the move with a longer >>>>>>>>search. >>>>>>> >>>>>>>1. Rxd4 cxd4 2. Re7+ Kb6 3. Qxd4+ Kxa5 4. b4+ Ka4 5. Ra7 Bb7 6. Rxb7 Qxd5 7. Rb6 >>>>>>>1-0 >>>>>> >>>>>>I always thought that simply 1. .. Kb6 was never refuted? >>>>> >>>>>You are absolutely right, missed that one. But I would like to see the engine >>>>>that plays 1..Kb6 >>>>> >>>>>Hmmm... >>>>> >>>>>Ed >>>> >>>> >>>>Here is one: (this on my dual xeon, searching around 2M nodes per second. WCCC >>>> opteron would be 8x faster... >>>> >>>> 12 13.94 -1.77 1. ... cxd4 2. Re7+ Kb6 3. Qxd4+ Kxa5 >>>> 4. Qc3+ b4 5. axb4+ Qxb4 6. Qc7+ Ka4 >>>> 7. b3+ Ka3 8. Qe5 Qb6 9. Qb2+ Kb4 10. >>>> c3+ Kc5 11. Qf2+ Kd6 12. Qxb6+ Kxe7 >>>> 13. c4 >>>> 12-> 20.05 -1.77 1. ... cxd4 2. Re7+ Kb6 3. Qxd4+ Kxa5 >>>> 4. Qc3+ b4 5. axb4+ Qxb4 6. Qc7+ Ka4 >>>> 7. b3+ Ka3 8. Qe5 Qb6 9. Qb2+ Kb4 10. >>>> c3+ Kc5 11. Qf2+ Kd6 12. Qxb6+ Kxe7 >>>> 13. c4 (s=2) >>>> 13 26.64 +1 1. ... cxd4 >>>> 13 34.01 -0.01 1. ... cxd4 2. Re7+ Kb6 3. Qxd4+ Kxa5 >>>> 4. b4+ Ka4 5. Qc3 Qxd5 6. Qc7 Qd1+ >>>> 7. Kb2 Qd4+ 8. Kb1 Qd1+ >>>> 13 2:46 -0.31 1. ... Kb6 2. Nb3 Bxd5 3. Qxd6+ Rxd6 >>>> 4. Rd2 Rhd8 5. Red1 Ne8 6. f4 Bf3 7. >>>> Rxd6+ Nxd6 8. Rd3 Be4 9. Rd2 a5 >>>> >>>>Plays Kb6 to avoid a perceived repetition after cxd4... >>> >>>Yep mine plays it too. New Challenge, which engine finds 5.Ra7!! >>> >>>This after: 1. Rxd4 cxd4 2. Re7+ Kb6 3. Qxd4+ Kxa5 4. b4+ Ka4 5. Ra7 Bb7 6. Rxb7 >>>Qxd5 7. Rb6 >>> >>>Ed >> >>Probably none. Is there a win after >> >>.... 7.Rb6 a5 8.Qe3 Rc8 9.Bxc8 Rxc8 10.Rxf6 Rxc2 11.Kxc2 Qa2+ 12.Kd3 Qc4+ 13.Kd2 >>Qa2+ 14.Ke1 Qa1+ 15.Kf2 Qxf6 >> >>? >>regards >>Andy > >8.Qe3 is wrong > >For the correct line see > >http://www.talkchess.com/forums/1/message.html?449661 If you still have doubts about the correctness of Ra7 then here is an analysis with yace after some learning from that tree New game, b2r3r/4Rp1p/p2q1np1/1p1P4/kP1Q4/P4PPB/2P4P/1K6 w - - 0 1 Analysis by Yace 0.99.87: 5.Qxf6 Qxf6 -+ (-12.47) Depth: 1 00:00:00 5.Rxf7 Nxd5 -+ (-6.54) Depth: 1 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 1 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 1 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 2 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 2 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 3 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 3 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 4 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 4 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 5 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 5 00:00:00 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 6/22 00:00:00 38kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 6/25 00:00:00 184kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 7/25 00:00:00 184kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 7/30 00:00:01 885kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 8/30 00:00:01 885kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 8/30 00:00:01 1102kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 9/30 00:00:01 1102kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 9/38 00:00:06 8142kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 10/38 00:00:06 8142kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 10/42 00:00:17 24994kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 11/42 00:00:17 24994kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 11/44 00:01:02 90981kN 5.Ra7 Bb7 6.Rxb7 Qxd5 7.Rb6 a5 8.Ra6 Ra8 9.Qe3 Rxa6 10.Kb2 +- (5.57) Depth: 12/44 00:01:02 90981kN (, 16.09.2005) It seems clear that white is winning based on that score. The only significant improvement to consider for black after 1...cxd4 is maybe 2...Kb8 instead 2...Kb6 because 3...Qc5 is not a big improvement and white has a winning advantage after 4.Qxf6+ Qd6 5.Be6!! Bxd5 6.b4 After that line analysis by yace gives +3.06 for white and yace can take the score back to the position after 2...Kb6 by going backward New game, 3r3r/4Rp1p/pk1qBQp1/Np1b4/1P6/P4PP1/2P4P/1K6 b - b3 0 1 Analysis by Yace 0.99.87: <snipped> 6...Ba8 7.Qxf7 Rb8 8.Rd7 Qf8 9.c4 bxc4 10.Rd6+ Qxd6 11.Nxc4+ Kc6 12.Nxd6 Kxd6 13.Bc4 a5 14.Qe6+ Kc7 +- (2.19) Depth: 12/43 00:01:57 156827kN 6...Ba8 7.Qxf7 Rb8 8.Rd7 Qf8 9.c4 bxc4 10.Rd6+ Qxd6 11.Nxc4+ Kc6 12.Nxd6 h5 13.f4 Rhg8 +- (2.59) Depth: 13/43 00:02:08 169994kN 6...Ba8 7.Qxf7 Rb8 8.Rd7 Qf8 9.c4 bxc4 10.Rd6+ Qxd6 11.Nxc4+ Kc6 12.Nxd6 Kxd6 13.Qd7+ Ke5 14.f4+ Ke4 15.Bc4 Rhd8 16.Qe6+ +- (3.06) Depth: 13/51 00:04:46 383590kN 6...Ba8 7.Qxf7 Rb8 8.Rd7 Qf8 9.c4 bxc4 10.Rd6+ Qxd6 11.Nxc4+ Kc6 12.Nxd6 Kxd6 13.Qd7+ Ke5 14.f4+ Ke4 15.Bc4 Rhd8 16.Qe6+ +- (3.06) Depth: 13/51 00:05:40 451103kN (, 16.09.2005) New game, b2r3r/4Rp1p/pk1q1np1/Np1P4/3p1Q2/P4PPB/1PP4P/1K6 w - - 0 1 <snipped> 3.Qxd4+ Qc5 4.Qxf6+ Qd6 5.Be6 Bxd5 6.b4 Ba8 7.Qxf7 +- (3.06) Depth: 12/43 00:00:28 42337kN (, 16.09.2005) Unfortunately 2...Kb8 also is not good because of 3.Qd4 when yace can see +3.14 for white after that line and the conclusion is learning a score of +3.06 for white after 1...cxd4 New game, bk1r3r/4Rp1p/p2q1np1/Np1P4/3Q4/P4PPB/1PP4P/1K6 b - - 0 1 <snipped> +- (3.14) Depth: 13/45 00:01:28 128741kN 3...Nd7 4.Bxd7 Bxd5 5.c4 Qxe7 6.Qb6+ Ka8 7.Qxa6+ Kb8 8.Qb6+ Ka8 9.Bc6+ Bxc6 10.Nxc6 Rd7 11.Nxe7 Rxe7 12.Qa6+ Ra7 13.Qxb5 +- (3.14) Depth: 13/45 00:01:44 147183kN (, 16.09.2005) New game, b2r3r/k4p1p/p2q1np1/Np1P4/3p1Q2/P4PPB/1PP4P/1K2R3 w - - 0 1 Analysis by Yace 0.99.87: <snipped> 2.Re7+ Kb6 3.Qxd4+ Qc5 4.Qxf6+ Qd6 5.Be6 Bxd5 6.b4 Ba8 7.Qxf7 +- (3.06) Depth: 12/34 00:00:06 9205kN 2.Re7+ Kb6 3.Qxd4+ Qc5 4.Qxf6+ Qd6 5.Be6 Bxd5 6.b4 Ba8 7.Qxf7 +- (3.06) Depth: 12/36 00:00:15 25090kN (, 16.09.2005) 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.