Author: James Robertson
Date: 23:19:49 03/31/00
Go up one level in this thread
On March 31, 2000 at 20:11:01, Rory wrote: >On March 31, 2000 at 02:29:31, Jeremiah Penery wrote: > >>This position came from Valentin Albillo's site, on one of the endgame pages. >> >>In this position: >> >>[D]8/8/8/8/3p4/2k1p3/2N1P3/3K4 w - - >> >>White is supposed to capture the first black pawn on ply 39 (move 20), and the >>second one shortly after. I ran my modified Crafty on this with all 5-man TBs >>to a depth of 45, with some strange results: >> >>White(1): 8/8/8/8/3p4/2k1p3/2N1P3/3K4/ w >>analyze.White(1): end-game phase >> clearing hash tables >> time surplus 0.00 time limit 30.00 (3:00) >> nss depth time score variation (1) >><It's not very interesting until about here.> >> 37-> 58:30 0.78 1. Ne1 Kb3 2. Nd3 Kc3 3. Ne5 Kb3 4. >> Nd3 >> 38 58:37 -- 1. Ne1 >> 38 58:41 ++ 1. Ne1!! >> 38 62:52 4.54 1. Ne1 Kb2 2. Nd3+ Kc3 3. Nc1 Kb2 4. >> Na2 Kb1 5. Nb4 Kb2 6. Nd5 Kb3 7. Nc7 >> Kc3 8. Nb5+ Kc4 9. Nd6+ Kb3 10. Ne4 >> Kb2 11. Nc5 Kc3 12. Ke1 Kb2 13. Ne6 >> Kc3 14. Nf4 Kc4 15. Kf1 Kc5 16. Kg2 >> Kd6 17. Nd3 Ke7 18. Kf3 Kf8 19. Ke4 >> Kg7 20. Kxd4 Kf6 21. Kxe3 >> >><This is what the analysis shows should happen, approximately. Looks good so >>far. We're into TBs now, but since it's too close to the leaf, Crafty doesn't >>probe them here, I think.> >> >> 38-> 62:52 4.54 1. Ne1 Kb2 2. Nd3+ Kc3 3. Nc1 Kb2 4. >> Na2 Kb1 5. Nb4 Kb2 6. Nd5 Kb3 7. Nc7 >> Kc3 8. Nb5+ Kc4 9. Nd6+ Kb3 10. Ne4 >> Kb2 11. Nc5 Kc3 12. Ke1 Kb2 13. Ne6 >> Kc3 14. Nf4 Kc4 15. Kf1 Kc5 16. Kg2 >> Kd6 17. Nd3 Ke7 18. Kf3 Kf8 19. Ke4 >> Kg7 20. Kxd4 Kf6 21. Kxe3 >> 39 63:01 ++ 1. Ne1!! >> 39-> 124:20 4.93 1. Ne1 Kb2 2. Nd3+ Kc3 3. Nc1 Kb2 4. >> Na2 Kb1 5. Nb4 Kb2 6. Nd5 Kb3 7. Nc7 >> Kc3 8. Nb5+ Kc4 9. Nd6+ Kb3 10. Ne4 >> Kb2 11. Nc5 Kc3 12. Ke1 Kb2 13. Ne6 >> Kc3 14. Nf4 Kc4 15. Kf1 Kc5 16. Kg2 >> Kd6 17. Nd3 Ke7 18. Kf3 Kf8 19. Ke4 >> Kg7 20. Kxd4 Kf6 21. Kxe3 >> 40 125:13 ++ 1. Ne1!! >> 40-> 417:29 5.32 1. Ne1 Kb2 2. Nd3+ Kc3 3. Nc1 Kb2 4. >> Na2 Kb1 5. Nb4 Kb2 6. Nd5 Kb3 7. Nc7 >> Kc3 8. Nb5+ Kc4 9. Nd6+ Kb3 10. Ne4 >> Kb2 11. Nc5 Kc3 12. Ke1 Kb2 13. Ne6 >> Kc3 14. Nf4 Kc4 15. Kf1 Kc5 16. Kg2 >> Kd6 17. Nd3 Ke7 18. Kf3 Kf8 19. Ke4 >> Kg7 20. Kxd4 Kf6 21. Kxe3 >> >><Still, looking good, as before.> >> >> 41 418:44 -- 1. Ne1 >> 41 421:21 0.00 1. Ne1 Kb2 2. Nd3+ Kc3 3. Nc5 Kc4 4. >> Ne4 d3 5. exd3+ <HT> >> >><What happened? The drawscore remains until...> >> >> 43-> 530:24 0.00 1. Ne1 Kb2 2. Nd3+ Kc3 3. Nc5 Kc4 4. >> Ne4 d3 5. exd3+ <HT> >> 44 548:34 ++ 1. Ne1!! >> 44-> 1233:04 0.39 1. Ne1 Kb2 2. Nd3+ Kc3 3. Nc5 Kc4 4. >> Ne4 d3 5. exd3+ <HT> >> 45 1234:35 ++ 1. Ne1!! >> >><Will the score go back up again? Will it find mate from the TBs?> >> >>I'm wondering if someone with a faster machine, with more memory (for TB cache, >>primarily - my HD was going nuts, practically grinding my whole machine to a >>halt) could run this until it sees mate. I'd like to know how far out it is. :) >> >>Thanks, >>Jeremiah > >Dear Jeremiah >I put this position into Fritz6a running on an AMD750 with 384mb of ram and with >all 4 piece endgames installed to the harddrive and it gave me Mate in 100 from >the position in your diagram >It did this in 1 minute 27 seconds!! >take care >Rory This is impossible. If black loses the first pawn after 20 moves, he will lose the second one very quickly also (_certainly_ in under 10 moves), and from that point a KNP vs K ending can be won in less than 20 moves. This is almost certainly less than mate in 50. James
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.