Author: leonid
Date: 16:02:02 07/11/01
Go up one level in this thread
On July 11, 2001 at 17:55:48, Christoph Fieberg wrote: >On July 11, 2001 at 17:36:36, leonid wrote: > >>On July 11, 2001 at 15:20:17, Heiner Marxen wrote: >> >>>On July 11, 2001 at 14:33:12, Christoph Fieberg wrote: >>> >>>>On July 11, 2001 at 12:22:15, Heiner Marxen wrote: >>>> >>>>>On July 11, 2001 at 11:45:17, Christoph Fieberg wrote: >>>>> >>>>>>I composed the following position: >>>>>>[D]r2qk2r/8/8/qqqqqqqq/QQQQQQQQ/8/8/R2QK2R b KQkq - 0 80 >>>>>> >>>>>>Who can construct a game to reach it? >>>>>> >>>>>>In my first attempt I needed 161 moves (last move was 81.Ke1). >>>> >>>>161 moves was meant as half-moves! (as you could see from the notion 81.Ke1) >>>>Sorry, that I did not express excatly. >>> >>>I see! Sorry, could have guessed that myself. >>>To avoid this sort of confusion, in the context of computer chess half-moves >>>are often called "plies". >>> >>>>> >>>>>If the white K has moved, its castling rights KQ should be removed! >>>>> >>>> >>>>In my game both kings had moved! Therefore no castling possible any more. >>> >>>Aha. In that case the same mate in 8 does follow, but the solution time >>>of Chest reduces from 330.54 secs to 296.10 secs. I didn't expect the >>>difference to be that large... Chest still can surprise its author :-) >> >> >>Hi! I found this "strange" position to be very much like my "normal" one. When >>you indicated that it contain mate in 8, I thought that, probably, it is more >>that coincidence. >> >>LLchess mate solver. 600Mhz. No hash. Castling regarded as existing. >> >>First mate for white side found for 8 moves deep in 35 seconds. >> >>Obviosly mate for black side in 8 moves also existe and was found. >> >>Cheers, >>Leonid. > >Hi Leonid, Hi! Since I am here, would like to say that my previous time was not indicated by what way it was found and this could give too good impression for no reason. Solution that I gave before was found by selective search. By pure brute force my program will find first solution in much longer time. Brute force - 3 min and 58 sec. Both solutions: White Queen goes from A4 to B5. >your position indeed inspired me to construct the above one. The main difference >is that the position above is possible from a real game (within 69 moves). > >Did you see my comment: http://www.icdchess.com/forums/1/message.shtml?179231 ? Yes. You said that indicated position is pure fantasy and can't be reached from real game. I agree with you 100%. There is nothing practical in my positions with exception that they permit to enjoy some wild mate search and make mate solver rock solid. After all, game is shortest way to escape "real", anyway. Cheers, Leonid. >What about my challenge? Can you reach the position with less than 69 moves? I was not looking into this but it could be interesting to see for some people. Very good question in itself. >Best regards, >Christoph > >> >> >> >> >>>>>>Who needs less moves? >>>>> >>>>>IIRC, Popeye can construct proof games. But more than 300 plies? >>>> >>>>As explained only 161 plies are necessary! Can Popeye cope with it? >>> >>>I have no idea, since I did not really use that feature. >>>I have studied its source code for a similar job, i.e. help mates, >>>which uses basically the same technique. The program calculates a good >>>minimum distance between source and destination position and uses this >>>to prune early. >>> >>>Whether this method is good enough for this problem... I just don't know. >>>You could try it yourself: Popeye is sort of free (GNU copyleft IIRC). >>> >>>Regards, >>>Heiner >>> >>> >>>>Regards, >>>>Christoph >>>> >>>> >>>>>I cannot imagine any program to proove the shortest proof game for such >>>>>a position... but then, what do I know? >>>>> >>>>>>Best regards, >>>>>>Christoph >>>>> >>>>>Regards, >>>>>Heiner
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.