Author: Tim Foden
Date: 10:48:41 12/05/03
Go up one level in this thread
On December 05, 2003 at 11:38:25, Tord Romstad wrote: >On December 04, 2003 at 13:37:47, Leen Ammeraal wrote: > >>Do you mean that most engines crash when dealing with >>this position? If so, I don't see why. My engine Queen does not. > >Gothmog also doesn't crash, but it needs more than 10 seconds to >complete a 1-ply search! I have been waiting several minutes for the second >iteration to complete. > >Perhaps positions like this could even be useful, in order to find ways to >reduce the number of qsearch nodes. > >Tord When Leonid was always posting positions like these as checkmate problems I changed GLC to have a hard q-search depth limit equal to the current iteration depth, or a minimum of 4 ply. It lets GLC get to a better depth at least. The other thing I did was order moves initially to put the checking moves first (if any). For this position... Game stage: Opening Current eval: -0.325 Ply Time Score Nodes Principal variation 2 0.030 +10.003 12308 Rxg8+ Qhxg8 2. Qxh6+ Qfxh6 3. Bxb8 Qxa5 4. Rxa5 Qxd1 5. Qbxd1 2 0.030 +10.003 14964 Rxg8+ Qhxg8 2. Qxh6+ Qfxh6 3. Bxb8 Qxa5 4. Rxa5 Qxd1 5. Qbxd1 3 0.150 +9.603 100941 Rxg8+ {--} Qfxg8 2. Rxh6 Q7xh6 3 0.170 +8.981 106082 Rxg8+ Qfxg8 2. Rxh6 Q7xh6 3. Qxe8 Rxa3 3 0.230 +8.981 134445 Rxg8+ Qfxg8 2. Rxh6 Q7xh6 3. Qxe8 Rxa3 4 0.290 +8.581 165026 Rxg8+ {--} Qfxg8 2. Rxh6 Rxb1+ 3. Kxb1 Q7xh6 4 0.461 +4.486 252559 Rxg8+ Qfxg8 2. Rxh6 Rxb1+ 3. Qaxb1 Q5xh6 4. Qxh6 Rxa3+ 4 1.122 +4.487 602809 Bxb8 Bxg1 2. Rxh6 Qfxh6 3. Qxh6 Rxa3 4 2.173 +5.070 1165753 Qxc8 Rxb1+ 2. Kxb1 Qxd1+ 3. Qaxd1 Qxc8 4. Qaxf8 Qgxf8 5. Qxh4 4 2.183 +5.070 1166659 Qxc8 Rxb1+ 2. Kxb1 Qxd1+ 3. Qaxd1 Qxc8 4. Qaxf8 Qgxf8 5. Qxh4 5 2.203 +4.670 1185528 Qxc8 {--} Rxb1+ 2. Kxb1 Qxd1+ 3. Qaxd1 Qxc8 4. Qaxf8 Qgxf8 5. Qxh4 Rxh4 5 3.195 +0.240 1952314 Qxc8 Rxb1+ 2. Qdxb1 Rxa3 3. Rxg8+ Qfxg8 4. Qxe8 Qgxe8 5 8.502 +2.450 5302137 Bxb8 Qhxc1 2. Rxg8+ Qhxg8 3. Qdxc1 Nb3+ 4. Qa4xb3 Qxa5 5. Qaxf8 Rxb3 6. Qaxa5 Rxb1+ 5 9.384 +4.532 5872111 Rxg8+ Qfxg8 2. Rxh6 Rxb1+ 3. Qcxb1 Q7xh6 4. Qf6+ Qdxf6 5. Nexf6 Qxe1 6. Qxh5 Qexh5 7. Qbxe1 5 12.628 +4.671 7990534 Rxh6 {++} Qfxh6 2. Rxg8+ Kxg8 3. Bxb8 Qhxc1 4. Qbxc1 5 13.780 +4.866 8635169 Rxh6 Rxb1+ 2. Qcxb1 Qfxh6 3. Qxd8 Qexd8 4. Qxh5 Rxa3 5. Rxg8+ Qdxg8 6. Qexh4 Rxa2+ 5 19.047 +4.866 12109k Rxh6 Rxb1+ 2. Qcxb1 Qfxh6 3. Qxd8 Qexd8 4. Qxh5 Rxa3 5. Rxg8+ Qdxg8 6. Qexh4 Rxa2+ 6 37.534 +4.489 21484k Rxh6 Rxb1+ 2. Kxb1 Qfxh6 3. Qxd8 Qexd8 4. Qxc8 Qxc8 5. Rxg8+ Kxg8 6. Qxh5 Rxa3 7. Qhxh4 Rxa2 8. Qxh6 Rxa4 Analyse>exit local: t=1:46.21 nps=571522.4 n=60703106 (2.1% / 97.9%) total: t=1:46.21 nps=571522.4 n=60703106 stats: fh=96.9%/1.14%/0.674% draws=10 trans: probes=1282039 hits=229767 (17.92%) draft=163348 (12.74%) tcuts: exact=45 (0.00%) upper=42875 (3.34%) lower=27595 (2.15%) tstor: exact=200 (0.02%) upper=879923 (82.52%) lower=186147 (17.46%) ehash: probes=0 hits=0 (-1.$%) ext: check=880569 recap=130417 ppush=0 1rep=2941 thrt=0 q-moves: gen=412196422 tested=56938639 made/un=49633427 max-dep=6 max eval diff: part-1=2.723 part-2=5.610 Cheers, Tim.
This page took 0.01 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.