Author: Richard Pijl
Date: 12:56:19 12/20/02
Go up one level in this thread
On December 20, 2002 at 13:13:27, Roy Eassa wrote: >On December 20, 2002 at 04:24:35, Richard Pijl wrote: > >>On December 19, 2002 at 19:39:56, Roy Eassa wrote: >> >>>White to move & win >>>[d] 7k/5r2/6PK/6P1/8/8/5P1R/8 w - - 0 1 >>> >>>7k/5r2/6PK/6P1/8/8/5P1R/8 w - - 0 1 >>> >>>1.g7+! >>> >>> [1.Kh5 Rf4 only draws] >>> >>>1...Kg8 >>> >>> [1...Rxg7 2.Rh5 Kg8 >>> (2...Rf7 3.Kg6+ Kg8 4.Rh8+ +-) >>> 3.g6 Rd7 4.Ra5 +-] >>> >>>2.Rh3! Rf3 3.Rg3! Rxf2 >>> >>> [3...Rf4 4.Re3 Rh4+ 5.Kg6 Re4 6.Kf5 Rc4 7.Re7 +-] >>> >>>4.Ra3 Rh2+ 5.Kg6 Ra2 6.Rf3 Ra6+ 7.Rf6 Rb6 8.Kh6! Rb8 9.Rf8+ +- >>> >>> [9.g6 also works] >> >>1.g7+ is found by the Baron on slow hardware (PII-500) in a few seconds and with >>a large score: >> >> ply time nodes score pv >> >> 19(31). 520 80021040 551 g6g7 h8g8 h2h3 f7f4 h3e3 f4h4 h6g6 h4e4 g6f5 >>e4d4 e3e7 g8h7 f2f4 d4d6 f5e5 d6g6 f4f5 g6g5 e5f4 > > >I don't think that it sees the best defense for Black. IMHO it's important to >see at least the first 5 ply of the solution in this case, because it represents >some very clever defense by Black and some additional critical but hard-to-find >moves by White. Anything less does not demonstrate understanding of the (only) >way to win the game from the starting position, against stiff defense. As is >often the case, many programs might select the correct first move or two for the >wrong reason. I understand that, but are you sure 2..Rf3 is absolutely strongest? Baron finds 2..Rf3 3.Rg3 quite fast, but then switches to 2..Rf4. When I force 2..Rf3 Baron has no trouble finding the remaining correct moves: (different version now, on Athlon MP 2000+, and with 5 stone TB, clean hash) ply time nodes score pv 4( 7)& 0 637 254 h3h1 f3f2 h6g6 f2f6 g6h5 4( 7)& 0 964 255 h3h4 f3f4 h6h5 f4f2 4( 7). 0 1095 255 h3h4 f3f4 h6h5 f4f2 5( 9)& 0 2036 260 h3h4 f3f4 h6h5 f4f2 h5h6 f2f4 5( 9)& 0 3338 277 h3g3 f3f2 g3e3 f2h2 h6g6 h2f2 5( 9). 0 3525 277 h3g3 f3f2 g3e3 f2h2 h6g6 h2f2 6(11)& 0 11478 251 h3h1 f3f2 h1e1 f2h2 h6g6 h2e2 e1g1 e2e1 6(11). 0 11665 251 h3h1 f3f2 h1e1 f2h2 h6g6 h2e2 e1g1 e2e1 7(12)& 0 18061 275 h3h1 f3f2 h1e1 f2h2 h6g6 h2e2 e1f1 e2f2 7(12). 0 21171 275 h3h1 f3f2 h1e1 f2h2 h6g6 h2e2 e1f1 e2f2 8(15)- 0 37429 245 h3h1 f3f2 8(16)& 0 60946 268 h3g3 f3f2 g3d3 f2h2 h6g6 h2d2 d3f3 d2d6 f3f6 8(16). 0 60946 268 h3g3 f3f2 g3d3 f2h2 h6g6 h2d2 d3f3 d2d6 f3f6 9(15)& 0 74265 271 h3g3 f3f2 g3e3 f2h2 h6g6 h2e2 e3f3 e2e6 f3f6 e6d6 g6h6 9(17). 0 81978 271 h3g3 f3f2 g3e3 f2h2 h6g6 h2e2 e3f3 e2e6 f3f6 10(19)+ 0 113207 301 h3g3 10(19)& 0 146124 300 h3g3 f3f2 10(19). 0 169744 300 h3g3 f3f2 11(21)& 0 283852 329 h3g3 f3f7 g3d3 f7d7 d3c3 d7d6 h6h5 d6d4 c3b3 g8g7 b3b7 g7f8 11(21). 0 316647 329 h3g3 f3f7 g3d3 f7d7 d3c3 d7d6 h6h5 d6d4 c3b3 g8g7 b3b7 g7f8 12(23)& 1 540456 329 h3g3 f3f7 g3d3 f7d7 d3c3 d7d6 h6h5 d6d4 h5g6 d4c4 c3f3 12(23). 1 585077 329 h3g3 f3f7 g3d3 f7d7 d3c3 d7d6 h6h5 d6d4 h5g6 d4c4 c3f3 13(23)+ 2 995844 359 h3g3 13(24)& 3 1373628 413 h3g3 f3f7 g3b3 f7b7 b3f3 b7g7 g5g6 g7d7 f3e3 d7d8 f2f4 g8f8 f4f5 13(24). 3 1428943 413 h3g3 f3f7 g3b3 f7b7 b3f3 b7g7 g5g6 g7d7 f3e3 d7d8 f2f4 g8f8 f4f5 14(28)+ 3 1766930 443 h3g3 14(30)& 5 2551509 442 h3g3 f3f7 14(30). 6 2816492 442 h3g3 f3f7 15(30)& 8 3744944 467 h3g3 f3f7 g3e3 f7e7 e3f3 e7g7 g5g6 g7d7 f3e3 d7d8 f2f4 g8f8 f4f5 15(30). 9 4167086 467 h3g3 f3f7 g3e3 f7e7 e3f3 e7g7 g5g6 g7d7 f3e3 d7d8 f2f4 g8f8 f4f5 16(33)+ 12 5381506 497 h3g3 16(33)& 14 6225338 511 h3g3 f3f4 g3e3 f4h4 h6g6 h4e4 g6f5 e4d4 e3e7 g8h7 f5e5 d4d8 f2f4 d8a8 f4f5 a8a5 e5e4 a5a4 e4e3 a4a5 16(33). 17 7202389 511 h3g3 f3f4 g3e3 f4h4 h6g6 h4e4 g6f5 e4d4 e3e7 g8h7 f5e5 d4d8 f2f4 d8a8 f4f5 a8a5 e5e4 a5a4 e4e3 a4a5 17(34)& 23 9706273 518 h3g3 f3f4 g3e3 f4h4 h6g6 h4e4 g6f5 e4d4 e3e7 d4d6 f2f4 d6b6 f5e5 b6b5 e5f6 17(34). 27 10765849 518 h3g3 f3f4 g3e3 f4h4 h6g6 h4e4 g6f5 e4d4 e3e7 d4d6 f2f4 d6b6 f5e5 b6b5 e5f6 > >With this problem I think seeing the first 5 ply may be enough; for some other >positions the programs must see even further to prove that they "understand" >what's going on. > >How long before your program shows this as the PV? : > >1.g7+ Kg8 2.Rh3 Rf3 3.Rg3
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.