Author: Duncan Roberts
Date: 12:08:11 02/13/06
Go up one level in this thread
On February 12, 2006 at 15:09:41, Marc Bourzutschky wrote: >On February 12, 2006 at 12:45:38, Duncan Roberts wrote: > >>On February 12, 2006 at 12:04:45, Marc Bourzutschky wrote: >> >>>On February 11, 2006 at 04:53:29, h.g.muller wrote: >>> >>>>Interesting. I did a lot of 4- and 5-men TBs including fairy pieces (mainly >>>>leapers) in a systematic effort to assign an accurate end-game value to pieces >>>>(in absense of pawns). It never occurred to me to use non-standard board-size, >>>>though. I guess the 10x8 board is of practical interest, due to >>>>Capablanca/Gothic Chess. >>>> >>>>Is your (or Marc's)work published somewhere? >>> >>>Vaclav Kotesovec has done a lot of work on fairy pieces, including leapers on >>>generalized boards: >>> >>>http://web.telecom.cz/Vaclav.kotesovec/ >>> >>>John Beasley and I looked at Q vs. R endings on generalized boards, because the >>>lack of an obvious systematic winning procedure on 8x8 suggested o John it may >>>not be a general win on larger boards. Indeed, it is only a general win on >>>boards up to 15x15, but on 16x16 it is in general a draw. John had also thought >>>it would be a draw on 8x12 (which is used in a variant called "Modern Courier >>>Chess") but the queen still wins there. This was published in the "British >>>Endgame Study News" in June and September 2004. >>> >>>Regarding leapers, John Beasley, Noam Elkies, and I have studied "doublet >>>leapers", which are pieces that combine two leapers into one piece. There is >>>only a small number of doublet leapers that can force mate, and only on board up >>>to 16x16. This was published in "Variant Chess" February 2005. >>> >>>We have also looked at "triplet leapers", combining three leapers into one. >>>Here one can construct cases that win on an arbitrarily large board. However, >>>it is only a finite set, and there are some triplets for which we are not sure >>>whether they are a general win or not. I have run tablebases on up to 90x90 for >>>some of these. >>> >>>-Marc >> >>what is your opinion on this post ? >>http://www.talkchess.com/forums/1/message.html?485845 >> >>duncn > >As someone else posted, Eiko Bleicher's "freezer" basically moves in that >direction. The vast majority of real life positions can be analyzed without >going to the trouble of traditional full information tablebases. However, there >always seem to remain a small number of positions where this will not work. >Most of these exceptional positions are endings that are draws in general, but >have some positions that are complicated wins. krnknn and I believe krrnkrr >belong to this category. > >-Marc as they complement each other do you have any interest in persuing both at the same time ? duncan
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.