Author: Tom Likens
Date: 08:40:47 07/15/04
Go up one level in this thread
On July 15, 2004 at 06:14:13, Eduard Nemeth wrote: Hello Eduard, As always, your posts are both entertaining and illustrative. I ran the seven endgame positions you provided against my engine and got mixed-results, (quite similar to Jonny). All the positions were run on an Athlon 1800+, in analysis mode, without tablebases. Some comments are interspersed below the diagrams. >Position 1 (1...Qxf3 and black wins): > >[D]6k1/1b3p2/1p2p1P1/pPp1P3/2P2Q1P/5PKB/8/3q4 b - - 0 1 Djinn also likes 1... Qxe1+ for a long time. After roughly 12 minutes it fails high and switches to 1 ... Qxf3. A failure in my opinion, since it would never find it at tournament time controls. >Position 2 ( b5-b4! Once move to win for black): > >[D]8/3r4/7P/kp4K1/8/1P6/P7/8 b - - 0 1 I really like this position. It's a good position for people as well as machines. The ideas are relatively straightforward, but difficult for most engines. Djinn didn't "understand" it until are I started walking it down the main line. Very nice. >Test 3 (a2-a4! Once move to win for white): > >[D]8/8/1ppk4/p4P2/4K3/8/PP6/8 w - - 0 1 Djinn doesn't get this, but I believe it should, since I have code that is supposed to handle these type of cases. I'll have a look tonight when I get home from work. >Test 4 (Only Kxa3 win): > >[D]8/8/k7/P7/1K6/n3B3/1P6/8 w - - 0 1 It gets this one fairly quickly, with a score similar to Jonny. >Position 5. Is this a hard position without TBs? >1. Bd2! wins, or draw. > >[D]8/8/6p1/6Kp/4kP2/8/8/4B3 w - - 0 1 Djinn also like 1. Bg3 with a high score +10 similar to Jonny. I'm puzzled Eduard, are you sure 1. Bg3 isn't a win for white? I'm curious about the refutation. >Position 6. Nullmove Test - Mate >1.Re1! >[D]8/8/p3R3/1p5p/1P5p/6rp/5K1p/7k w - - 0 1 It gets this immediately (less than 1 second), which is gratifying considering all the zugzwang, null-move code I've added. >Text Position after clear hash – important - : > >Position 7. Mate in 2: > >[D]8/4B3/5K1p/5p1k/5Pb1/5p1p/5P1P/8 w - - 0 1 It also gets this very quickly (again less than a second). Once again, thanks for the excellent positions- please keep them coming, and I'm still curious about 1 Bg3 on position #5. best regards, --tom >Eduard
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.