Author: Peter Kasinski
Date: 10:02:36 12/20/05
Go up one level in this thread
On December 20, 2005 at 03:51:12, Yuriy Lyapko wrote: >To solve such positions engine must have built-in knowledge about king blocade. >I don't know about new engines but one year ago there were only 2 engines with >such a knowledge(here are results on Athlon 1700+): > >Patzer 3.61 needs ~6 seconds to show full PV of solution: > >no 1 337 -551 4 Rc8 >no 1 338 +321 90 Rh8+ >no 2 373 +321 99 Rh8+ Kg4 >no 3 390 +281 430 Rh8+ Kg4 >no 3 407 -550 1081 Rh8+ Kg4 >yes 3 408 -26 1329 Rb1 cxb1=Q Bxb1 >yes 4 428 -28 1621 Rb1 cxb1=Q Bxb1 g4 >yes 5 429 -34 3461 Rb1 Bh3+ Kg1 >yes 6 446 -82 7287 Rb1 cxb1=Q >yes 6 484 -457 13638 Rb1 cxb1=Q Bxb1 e3 >no 6 557 -253 15252 Rh8+ Kg4 h3+ Kf4 Rf8 c1=Q Be6 Ke5 >no 7 629 -253 21717 Rh8+ Kg4 h3+ Kf4 Rf8 c1=Q Be6 Ke5 Bxf5 >no 8 789 -260 42230 Rh8+ Kg4 h3+ Kf4 Rf8 c1=Q Be6 Ke5 Bxf5 > Qd2 >no 9 1253 -257 94837 Rh8+ Kg4 Rf8 c1=Q >no 10 1544 -257 133822 Rh8+ Kg4 Rf8 c1=Q h3+ Kf4 >yes 10 1672 -210 148623 Rb1 >yes 11 4484 -12 439543 Rb1 cxb1=Q Bxb1 e3 Bxf5 e2 f3 e1=Q Bg4 > Qc3 >yes 12 5249 -37 543342 Rb1 cxb1=Q Bxb1 e3 Bxf5 e2 Bg4 e1=Q h3 > Qd2 Kf3 Qd3+ Kg2 >yes 13 9899 -40 816985 Rb1 cxb1=Q Bxb1 e3 Bxf5 e2 Bg4 e1=R h3 >yes 14 16601 -42 1256913 Rb1 cxb1=Q Bxb1 e3 Bxf5 e2 Bg4 e1=Q h3 > >Result: Success (1 of 1 solved so far - 100%) >Found in: 1672 ms (00:00:01.672) > >my Bestia needs 35 seconds to switch from losing 4.f3? to 4.Bg4! > >no 2 0 +358 313 b8h8 h4g4 >no 2 0 +158 511 b8h8 h4g4 >yes 2 0 -37 851 b8b1 c2b1Q >yes 3 10 -47 1697 b8b1 f5h3 g2g1 c2b1Q a2b1 h3f5 >yes 4 50 -47 4371 b8b1 f5h3 g2g1 c2b1Q >yes 5 50 -37 5258 b8b1 c2b1Q a2b1 h4g4 b1c2 >yes 5 60 -37 8730 b8b1 c2b1Q a2b1 h4g4 b1c2 >yes 6 60 -46 10647 b8b1 c2b1Q a2b1 h4g4 h2h3 g4f4 b1c2 >yes 6 110 -46 24034 b8b1 c2b1Q a2b1 h4g4 h2h3 g4f4 b1c2 >yes 7 230 -71 25777 b8b1 c2b1Q >yes 7 460 -71 62193 b8b1 c2b1Q >yes 7 460 -271 62694 b8b1 c2b1Q >no 7 780 -270 100889 b8h8 >no 7 870 -238 107350 b8h8 h4g4 h8f8 c2c1Q h2h3 g4f4 a2e6 > f4e5 e6f5 >no 7 1110 -238 146290 b8h8 h4g4 h8f8 c2c1Q h2h3 g4f4 a2e6 > f4e5 e6f5 >no 8 1280 -238 169856 b8h8 h4g4 h8f8 c2c1Q h2h3 g4f4 a2e6 > f4e5 e6f5 c1b1 >no 8 1470 -238 212836 b8h8 h4g4 h8f8 c2c1Q h2h3 g4f4 a2e6 > f4e5 e6f5 c1b1 >no 9 2720 -228 414942 b8h8 h4g4 h8f8 c2c1Q h2h3 g4f4 a2e6 > f4e5 e6f5 c1c7 g2g3 >yes 9 3110 -227 492352 b8b1 >yes 9 3280 -213 521292 b8b1 >yes 9 3330 -213 521292 b8b1 >yes 9 3640 -39 563434 b8b1 c2b1Q a2b1 e4e3 b1f5 e3e2 g2f3 > e2e1Q f5g4 >yes 9 4040 -39 621359 b8b1 c2b1Q a2b1 e4e3 b1f5 e3e2 g2f3 > e2e1Q f5g4 >yes 10 4170 -64 635797 b8b1 c2b1Q >yes 10 5940 -64 969853 b8b1 c2b1Q >yes 10 6160 -102 1011739 b8b1 c2b1Q a2b1 e4e3 b1f5 e3e2 f5g4 > e2e1N g2f1 h4g4 f1e1 g4f3 e1d2 g5g4 >yes 10 7990 -102 1349590 b8b1 c2b1Q a2b1 e4e3 b1f5 e3e2 f5g4 > e2e1N g2f1 h4g4 f1e1 g4f3 e1d2 g5g4 >yes 11 8860 -77 1491340 b8b1 >yes 11 8860 -77 1491340 b8b1 >yes 11 9760 -41 1649349 b8b1 c2b1Q a2b1 e4e3 b1f5 e3e2 f2f3 > e2e1Q f5g4 e1e2 g2g1 e2a2 >yes 11 13990 -41 2481242 b8b1 c2b1Q a2b1 e4e3 b1f5 e3e2 f2f3 > e2e1Q f5g4 e1e2 g2g1 e2a2 >yes 12 15250 -41 2720604 b8b1 c2b1Q a2b1 e4e3 b1f5 e3e2 f2f3 > e2e1Q f5g4 e1e2 g2g1 e2a2 g1h1 >yes 12 30660 -41 5137399 b8b1 c2b1Q a2b1 e4e3 b1f5 e3e2 f2f3 > e2e1Q f5g4 e1e2 g2g1 e2a2 g1h1 >yes 13 34880 -46 5818892 b8b1 c2b1Q a2b1 e4e3 b1f5 e3e2 f5g4 > e2e1N g2f1 h4g4 f1e1 g4h3 e1e2 h3g2 > h2h3 > >Result: Success (1 of 1 solved so far - 100%) >Found in: 3110 ms (00:00:03.110) > >I wonder why other engines doesn't implement this knowledge which is rather >simple: >1.Stronger side has only one piece. >2.Its king has no moves. >It will help in 99% cases. > >Here are some more examples: > >[D]1NK1k2N/4p3/8/8/8/4p3/8/8 w - - > >1.Na6 e2 2.Nc7+ Kf8 3.Ne6+ Kg8 4.Nf7! e1=Q 5.Nfg5= > >[D]1N6/b7/8/7p/3p3k/5P2/6PK/8 w - - > >1.Nc6 d3 2.Nxa7 d2 3.Nb5 d1=Q 4.Nc3 Qd6+ 5.Kh1 Q~ 6.Ne4= > >[D]5k2/1q5p/8/5N1P/6PB/7K/8/8 w - - > >1.Be7+ Kg8 2.Nh6+ Kg7 3.g5! Qf3+ 4.Kh4 Qg2 5.Bd6! Qh1+ 6.Kg4 Qd1+ 7.Kh4 Qf3 >8.Be7= > >[D]5B1k/7p/6P1/4NnP1/8/8/K4p2/8 w - - > >1.g7+ Nxg7 2.Nf7+ Kg8 3.Bc5! f1=Q 4.Nh6+ Kh8 5.Bd6! ~ 6.Be5!= > >[D]3K4/3Rp1k1/7p/1r6/7P/6P1/2p5/8 w - - > >1.Rc7! Rb8+ 2.Kxe7 Rb7 3.Rxb7 c1=Q 4.Ke6+ Kg6 5.h5+ Kxh5 6.Rg7 Qf1 7.Ke7!= > >[D]3kn3/4pp2/n7/5B2/3N4/b6N/4K3/8 w - - > >1.Ng5 Nd6 2.Nxf7+ Nxf7 3.Ne6+ Ke8 4.Bg6! Bc1 5.Bd3! Nb8 6.Bb5+ Nd7 7.Kd1! B~ >8.Be2 Nf~ 9.Bh5+ Nf7 10.Be2 Nd~ 11.Bb5+ Nd7 12.Be2= > >[D]8/7k/3b1K1B/8/p7/p7/R7/1b6 w - - > >1.Ra1 a2 2.Ke6! Ba3! 3.Bf4! Bb2 4.Be5 a3 5.Kd5! Bg6 6.Bd4 Bf7+ 7.Ke4 Bc4 8.Rg1!= > >It is funny too see some great engines struggling even on endpositions of some >of these studies, but... >Rybka shows us that endgame knowledge doesn't equal to playing strength. > >Best regards, >George Thank you for your answer and interesting positions! PK
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.