Author: Paul
Date: 07:38:44 02/05/01
Go up one level in this thread
On February 05, 2001 at 02:27:12, Steffen Jakob wrote: >On February 03, 2001 at 20:03:33, Eduard Nemeth wrote: > >>It is a simple testposition ( my new composition :), but i think it is for >>chess-programs hard! :) >> >>[D]qrk5/1p2Kp2/pBp1pPb1/P1P1P3/8/8/8/8 b - - 0 1 >> >>Black win! > >Very nice position! > >Normally Hossa has no chance to find the solution. Some time ago I worked on an >idea to detect and evaluate such blocked positions. The main idea is to not only >evaluate a position statically but also to consider how the evaluation >progresses from ply to ply. If you already seem to have such a big advantage >like black in Eduards test position than you "should" be able to make this >advantage bigger within "some" plies. If you are not able to do this, then this >is a strong indication for a blocked position. I have written a liitle bit about >this on Hossa's homepage (http://www.jakob.at/steffen/hossa.html). > >I even have written some experimental code for this idea and it works partly. It >has some very nasty side effects though, and therefore it's not part of the >standard Hossa. But: if I activate this code then Hossa finds the solution Qa7 >after 16 seconds: > > depth value time nodes pv > 1 +15.82 0.02 3 1... g6e4 2.e7xf7 > 7/1 +17.33 0.03 15 1... g6h5 > <1> +17.33 0.03 17 1... g6h5 > 2 +17.49 0.03 27 1... g6h5 2.e7d6 > <2> +17.49 0.03 41 1... g6h5 2.e7d6 > 3 +17.70 0.03 85 1... g6h5 2.e7d6 h5f3 > <3> +17.70 0.03 189 1... g6h5 2.e7d6 h5f3 >? 4 +16.46 0.05 533 1... g6h5 2.b6c7 h5g6 3.c7xb8 c8xb8 >? <4> +16.46 0.06 691 1... g6h5 2.b6c7 h5g6 3.c7xb8 c8xb8 > 5 +16.45 0.07 1019 1... g6h5 2.b6c7 a8a7 3.c7xb8 a7xb8 > <5> +16.45 0.13 2349 1... g6h5 2.b6c7 a8a7 3.c7xb8 a7xb8 > 2/6 +16.16 0.16 3367 1... g6d3 2.e7e8 d3e4 3.e8xf7 c8d7 > 4.f7g7 > <6> +16.16 0.20 4326 1... g6d3 2.e7e8 d3e4 3.e8xf7 c8d7 > 4.f7g7 >! 7 +16.78 0.25 5893 1... g6d3 2.e7e8 d3e4 3.e8xf7 c8d7 > 4.f7g7 b8g8 5.g7h6 >! <7> +16.78 0.36 9796 1... g6d3 2.e7e8 d3e4 3.e8xf7 c8d7 > 4.f7g7 b8g8 5.g7h6 >? 8 +13.49 0.56 19307 1... g6d3 2.e7e8 d3e2 3.e8e7 e2h5 > 4.b6c7 b6 5.c7d6 b6xa5 6.d6xb8 >? <8> +13.49 0.64 22453 1... g6d3 2.e7e8 d3e2 3.e8e7 e2h5 > 4.b6c7 b6 5.c7d6 b6xa5 6.d6xb8 >! 9 +13.58 0.73 27196 1... g6d3 2.e7e8 d3e2 3.e8e7 e2d1 > 4.e7d6 d1c2 5.b6c7 b6 6.c7xb8 b6xa5 >! <9> +13.58 0.89 34857 1... g6d3 2.e7e8 d3e2 3.e8e7 e2d1 > 4.e7d6 d1c2 5.b6c7 b6 6.c7xb8 b6xa5 >? 10 +6.92 1.45 66720 1... g6d3 2.e7e8 d3c2 3.e8e7 c2e4 > 4.e7e8 a8a7 5.b6xa7 b8a8 6.a7b6 >? <10> +6.92 1.69 82593 1... g6d3 2.e7e8 d3c2 3.e8e7 c2e4 > 4.e7e8 a8a7 5.b6xa7 b8a8 6.a7b6 >! 11 +6.90 1.98 100809 1... g6d3 2.e7e8 d3f1 3.e8e7 f1e2 > 4.e7e8 e2d3 >! <11> +6.90 2.71 146194 1... g6d3 2.e7e8 d3f1 3.e8e7 f1e2 > 4.e7e8 e2d3 >! 12 +6.90 3.27 186272 1... g6d3 2.e7e8 d3f1 3.e8e7 f1e2 > 4.e7e8 e2d3 5.e8e7 d3e4 >! <12> +6.90 4.40 265299 1... g6d3 2.e7e8 d3f1 3.e8e7 f1e2 > 4.e7e8 e2d3 5.e8e7 d3e4 >! 13 +6.90 5.58 348430 1... g6d3 2.e7e8 d3e2 3.e8e7 e2f1 > 4.e7e8 f1c4 5.e8e7 c4e2 >!2/13 +7.38 15.98 998322 1... a8a7 2.b6xa7 b5 3.c5xb6 c5 > 4.e7d6 c4 5.b7 c8xb7 6.a7d4 b8d8 > 7.d6c5 g6d3 8.d4c3 d8d5 9.c5b4 >! <13> +7.38 16.36 1023359 1... a8a7 2.b6xa7 b5 3.c5xb6 c5 > 4.e7d6 c4 5.b7 c8xb7 6.a7d4 b8d8 > 7.d6c5 g6d3 8.d4c3 d8d5 9.c5b4 > >Best wishes, >Steffen. Hi! I've been thinking along the same lines yesterday when I saw this problem, but haven't implemented anything yet. It's great to see you've managed to get this working! Congratulations!! Now I'm going to be mean :) Does it also work on these positions? At least now you have some more test positions ;) [D]4k3/8/1p1p4/pPpPp1p1/P1P1PpPp/5P1P/2BB4/K7 w - - id White Wins; bm Bxf4 [D]5rk1/5p2/pr2pPp1/Pp1pP1Pp/1PpP3P/K1P5/8/8 w - - id Draw; am axb6 [D]8/8/8/1k3p2/p1p1pPp1/PpPpP1Pp/1P1P3P/QNK2NRR w - - id WW; bm Kd1 [D]3k2r1/5b2/2p5/rpPp1p1p/pP1PpPpP/P3P1P1/3K4/8 w - - id Draw; am bxa5 [D]8/3k3p/5p2/p1p1pPpP/PpPpP1P1/RP1P4/8/5K1R w - - id WW; bm h6 Best wishes, Paul
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.