Author: leonid
Date: 08:13:26 03/09/02
Go up one level in this thread
On March 09, 2002 at 10:16:04, Uri Blass wrote: >On March 09, 2002 at 09:46:43, leonid wrote: > >>On March 07, 2002 at 19:48:08, Slater Wold wrote: >> >>>On March 07, 2002 at 08:05:18, Art Basham wrote: >>> >>>>Some programs that I have miss this mate in 4..! >>>> >>>>1. Qc6 wins for white!! >>>> >>>>[D]2kr1b1r/1pp1qp1p/3p2p1/1P1Pp3/2Q1P1n1/R4N2/1PP2PPP/5RK1 w - - >>> >>>White(2): setboard 2kr1b1r/1pp1qp1p/3p2p1/1P1Pp3/2Q1P1n1/R4N2/1PP2PPP/5RK1 w - - >>> >>>White(1): st 180 >>>search time set to 180.00. >>>White(1): go >>> clearing hash tables >>>crafty is behind 32.0 on time, reducing by 1/96. >>> time surplus 0.00 time limit 3:00 (3:00) >>> nss depth time score variation (1) >>> 5 0.09 0.61 1. b6 Bh6 2. Ra8+ Kd7 3. Qb5+ c6 4. >>> dxc6+ Ke6 5. Qd5+ Kf6 6. cxb7 Kg7 >>> 5-> 0.09 0.61 1. b6 Bh6 2. Ra8+ Kd7 3. Qb5+ c6 4. >>> dxc6+ Ke6 5. Qd5+ Kf6 6. cxb7 Kg7 >>> 6 0.15 0.84 1. b6 Bg7 2. Ra8+ Kd7 3. Qb5+ c6 4. >>> dxc6+ Ke6 5. Qd5+ Kf6 6. cxb7 Rb8 7. >>> Ra7 >>> 6-> 0.18 0.84 1. b6 Bg7 2. Ra8+ Kd7 3. Qb5+ c6 4. >>> dxc6+ Ke6 5. Qd5+ Kf6 6. cxb7 Rb8 7. >>> Ra7 >>> 7 0.50 0.77 1. b6 Bh6 2. Rfa1 Rhf8 3. Ra8+ Kd7 >>> 4. Qb5+ c6 5. dxc6+ Ke6 6. Qd5+ Kf6 >>> 7. cxb7 Kg7 >>> 7 0.61 ++ 1. Qc6!! >>> 7 0.62 Mat04 1. Qc6 bxc6 2. dxc6 Kb8 3. Rfa1 Nf6 >>> 4. Ra8# >>> (3) 7-> 0.62 Mat04 1. Qc6 bxc6 2. dxc6 Kb8 3. Rfa1 Nf6 >>> 4. Ra8# >>> time=0.84 cpu=148% mat=-2 n=918801 fh=94% nps=1093k >>> ext-> chk=38957 cap=7162 pp=6426 1rep=5746 mate=731 >>> predicted=0 nodes=918801 evals=98158 >>> endgame tablebase-> probes done=0 successful=0 >>> hashing-> trans/ref=44% pawn=98% used=3% >>> SMP-> split=141 stop=14 data=6/32 cpu=1.25 elap=0.84 >>> >>>mate in 4 moves. >> >>I looked by curiousity this position with my mate solver and found that finding >>of mate goes slowly. Even when mate was found by few versions of selective >>search time was still identical. >> >>LLchess mate solver. Celeron 600Mhz. No hash. >> >>Mate in 4 moves found and took 0.21 sec. First move goes by white queen C4-C6. >> >>Initial move was real surprise! >> >>Leonid. > >Not a real surprise > >This move threats mate in 1 >There are not many moves that threat mate in 1 so it should be easy for >chess programs to find it very fast(all you need to do is extending when you see >a threat of mate in 1). Easy mate is the mate that start with checking move. This is not. Even so, mate in 4, with actual number of pieces, should take less time. To make this kind of mate position more quickly solvable, I must write special additional logic for specialized plys in my solver. Until now I never found time and will to do so. Sometime, I really have seen very slow 4 moves deep positions but only for fantasy mates. Those mates (not in finding but for searching for mate 4 moves deep) took more that 30 seconds. Some idea of what those positions are you can find by looking into usual, "One mate to solve". Leonid. >Uri
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.