Author: Heiner Marxen
Date: 14:39:18 11/04/01
Go up one level in this thread
On November 04, 2001 at 10:45:12, leonid wrote: >On November 04, 2001 at 08:59:37, Paul wrote: > >>On November 04, 2001 at 07:54:58, leonid wrote: >> >>>Hello! >> >>Hi ... Pretz's first thoughts on a p3/1000 ... >> >>>This one is easy: >>> >>>[D]Q7/3qq3/2qQqk2/1nQRNqq1/nBKRRNq1/PrbBQq2/1PrPq3/2Bb4 w - - >> >>00:05 WM9 07 Qh8+ Qgg7 Nxd7+ Qcxd7 Nh5+ Qfxh5 Rxe6+ Qgxe6 Qef4+ Qxf4 Rxf4+ Qf5 >>Rxf5+ Kg6 Qg1+ Qg5 Qxg5# Hi to both of you! 5 seconds? Well, Chest is not _that_ fast, but this one is comparatively easy: in 111.46 seconds (K7/600, 350 MB) Chest confirms that Qh8+ is the only solution in 9 moves: PV: Qh8+ Qgg7 Nxd7+ Qcxd7 Nh5+ Qfxh5 Rxe6+ Qgxe6 Qef4+ Qxf4 Rxf4+ Qf5 Rxf5+ Kg6 Qg1+ Qg4+ Rf4# Except for the last 2 plies identical to your line. Timing: # 1 0.00s 0kN 0.92 1- 0 # 2 0.00s 0kN [ 48.00] 1.00 1- 0 # 3 0.01s 0kN [ 8.92] 0.95 43- 0 # 4 0.08s [ 8.00] 3kN [ 7.19] 1.17 327- 0 # 5 0.42s [ 5.25] 21kN [ 6.79] 1.30 2391- 0 # 6 1.81s [ 4.31] 88kN [ 4.20] 1.56 10849- 0 # 7 7.28s [ 4.02] 357kN [ 4.07] 1.98 45853- 0 # 8 28.55s [ 3.92] 1406kN [ 3.94] 2.46 180809- 0 # 9 111.45s [ 3.90] 5597kN [ 3.98] 3.17 706062- 0 Yes, a very "mild" EBF. >>>But if you want to solve something that is not difficult, almost the same but >>>demand more thinking, then try this: >>> >>>[D]Q7/3qq3/2qNqk2/1nNRNqq1/nPKPRNq1/PrbBQq2/1PrQq3/2Bb4 w - - >> >>00:35 WM13 09 Qh8+ Qeg7 Nxd7+ Qexd7 Nh5+ Q5xh5 Rxf5+ Qfxf5 Qeh6+ Qhg6 Q6xg7+ >>Qgxg7 Qdh6+ Qfg6 Rf4+ Qxf4 Qxg6+ Ke7 Qgxg7+ Qf7+ Qxf7+ Kxd6 Qf4+ Ke7 Qhf8# >> >>Not sure of course whether these are the shortest possible mates ... This one is significantly harder for Chest. After 4.2 hours (K7/600, 350 MB) it confirms, again: Qh8+ is the only shortest solution in 13 moves: PV: Qh8+ Qeg7 Nxd7+ Qexd7 Nh5+ Q5xh5 Rxf5+ Qfxf5 Qeh6+ Qhg6 Q6xg7+ Qgxg7 Qdh6+ Qfg6 Rf4+ Qxf4 Qxg6+ Ke7 Qgxg7+ Qf7+ Qxf7+ Kxd6 Bf4+ Qe5 Bxe5# # 1 0.00s 0kN 0.87 1- 0 # 2 0.00s 0kN 1.00 1- 0 # 3 0.01s 0kN [ 7.98] 0.94 44- 0 # 4 0.06s [ 6.00] 3kN [ 7.39] 1.05 249- 0 # 5 0.21s [ 3.50] 12kN [ 4.79] 1.27 1171- 0 # 6 0.87s [ 4.14] 51kN [ 4.14] 1.64 4834- 0 # 7 3.53s [ 4.06] 219kN [ 4.25] 2.15 20696- 0 # 8 10.65s [ 3.02] 657kN [ 3.00] 2.95 61072- 0 # 9 38.86s [ 3.65] 2677kN [ 4.07] 3.91 217967- 0 # 10 171.46s [ 4.41] 11033kN [ 4.12] 4.87 1045649- 1 # 11 554.99s [ 3.24] 30770kN [ 2.79] 5.67 3823600- 4947 # 12 2482.73s [ 4.47] 112283kN [ 3.65] 5.53 18494015- 9748454 # 13 14165.08s [ 5.71] 558147kN [ 4.97] 5.27 111698871- 102950970 It starts out even faster than for the first problem, and although the EBF stays quite moderate, the additional 4 moves depth make a lot of difference. One statistics came as a surprise: for several depthes the EBF was not from enumerating alternate white moves, but from enumerating more black defense moves! I do not recall to have seen that before: depth w moves b moves w quot b quot mvx 13: 22 28 [ 22.000 1.273] 21 mvx 12: 167 327 [ 5.964 1.958] 109 mvx 11: 1550 5746 [ 4.740 3.707] 318 4 mvx 10: 6634 25390 [ 1.155 3.827] 1124 27 mvx 9: 24984 94753 [ 0.984 3.793] 3569 173 mvx 8: 89914 322532 [ 0.949 3.587] 14073 647 mvx 7: 364201 1143166 [ 1.129 3.139] 44813 2043 mvx 6: 1553495 3780853 [ 1.359 2.434] 171837 6558 mvx 5: 6995183 13062507 [ 1.850 1.867] 821338 22892 mvx 4: 32190163 47679286 [ 2.464 1.481] 2308374 86151 mvx 3: 137230996 151136869 [ 2.878 1.101] mvx 2: 128053285 32693591 [ 0.847 0.255] mvx 1: 1690987 0 [ 0.052 ] Between depth 10 and 5 (depth to go) black contributed more than white to EBF. Hmmm... isn't it already a long time that Pretz did not find a shortest solution? Can that happen at all? Shouldn't we replace Chest with Pretz? It is so much faster :-)) [ok, just kidding :-] Cheers, Heiner >>>Please indicate your result. > >Hi, Paul! > >Now you reached me for sure. On first and second position, your time is better >that mine. First took 9 second for selective search 9 moves deep, and second 62 >sec for selective in 13 moves. Both solutions are the shortest possible. I >solved both by brute force. Branching factor is very mild for both of them. > >Cheers, >Leonid.
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.