Author: James T. Walker
Date: 10:28:18 02/02/01
Go up one level in this thread
On February 02, 2001 at 12:39:31, Timothy J. Frohlick wrote: >On February 02, 2001 at 12:10:58, James T. Walker wrote: > >>On February 02, 2001 at 03:19:14, Timothy J. Frohlick wrote: >> >>>On February 02, 2001 at 01:28:59, Jouni Uski wrote: >>> >>>>[D]8/6kn/3B3p/5K1B/8/8/8/8 b - - >>>> >>>>resign 1-0! Why? Is this really white's win? >>>> >>>>JouniDate: 1/2/2001 >>> >>>Jouni, >>> >>>Gambit Tiger without tablebases solves this as a mate in 46 in 26 minutes on a >>>PII 333 with 48 Megs Hash. >>> >>>1... Ng5 2. Bc5 Nf7 3. Bd4+ Kf8 4. Kf6 Ng5 5. Bc5+ Kg8 6. Kf5 Nf7 7. Be7 Ng5 8. >>>Bb4 Nf7 9. Kg6 Ne5+ 10. Kf6 Nd7+ 11. Ke7 Ne5 12. Bc3 Nc6+ 13. Ke8 Kh7 14. Kf7 >>>Ne5+ 15. Kf6 Nc6 16. Bf3 Nd8 17. Bb4 h5 18. Be4+ Kh8 19. Be7 Nc6 20. Bxc6 Kh7 >>>21. Be4+ Kg8 22. Kg6 Kh8 23. Bd5 h4 24. Bf6# >>> >>> >>>1-0 >>> >>>Maybe tablebases are not all that great after all. They don't always find the >>>shortest solution to a problem. >>> >>>Tim Frohlick >> >>Why mate in 46 and then you give a mate in 24 line? I believe the "mate in 46" >>is true as Shredder 5 finds a mate in 47 (maybe a difference in terms) but where >>did the above line which is 24 moves come from? >>Jim > > >Jim, > >Mate in 46 ply or 23 moves that is. I do believe that this is the shortest >mate. It sounds better to say the bigger number even though that is not how you >claim "mate in". What I did to solve the above problem was to let GT run out to >14 or 15 ply seaarches and hand-pick the end move. It seems that the "horizon >effect" can be partially eliminated by this method. In addition, I think that >GT would not even get to a 46 ply search in a week of searching on a 333 Pentium >II. The intelligent way to solve these long mate problems without a tablebase >is to do what I have done. > >It is not "Green" ie environmentally friendly to let one's computer run for days >just to solve one chess problem.. N'est pas? > >Tim "Save the Electrons" Frohlick Hello Tim, Well I don't believe the mate in 23/24 moves. I seriously doubt that Tiger could mate any of the top programs in 23 moves and I think the tablebases are the best method to find mate here. It is interesting to me that the different programs all using the same tablebases find different distances to mate here. Although the shortest was mate in 46/47 some found mate in 74 (Fritz 6)and some found mate in 49 (Junior 6) and all used the same tablebases to find these scores. Regards, Jim
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.