Since we published on CCM the claim of Lutz Neweklowsky, we received a few responses from readers and other involved people. So:
18st of April 2000
Joost de Heer pointed to us the page with mate in 262 under No. 60.
Of course, it wasn't clear whether the position was meant to be mate in 262 with dualfree longest variation, so I wrote 25th of April 2000 to
Ken Thompson, one of the authors the following:
Dear mr. Thompson, I am writing you as you are suppossed to be one of the authors of the longest known moremover. What's the matter? Mr. Bedrich Formanek, president of Permanent Comission for Chess Composition of FIDE (in fact the head of all organized chess composition) received the claim from mr. Lutz Neweklowsky of Germany that he composed new record. As I know mr. Formanek personally from our local meetings and I am running extensive site about chess composition, he asked me to publish Neweklowsky's claim to give wider community chance to react. You can read about it on my page http://members.tripod.com/~JurajLorinc/chess/newek258.htm Later mr. Formanek received the mail from Joost de Heer, telling: "On http://www.xs4all.nl/~timkr/chess2/diary.htm you will find a mate in 262! This is an improvement over Neweklowsky's 'record'." I checked it and yes, it seems to be the record. The question is the following: is it true that the longest (main) variation has no duals? Dual in this case means the chance on n-th move for white to mate in 263-n (or less) moves in two or more different ways. As you have database of the positions I think it might be possible to prove that main variation has no duals. Another question - where was your problem published for the first time? Somewhere in print, or in mailing list or on some www page? This is important as the primary source is usually published with diagram as the names of authors are too. Best wishes, JUraj.
Dear JU. Lorinc, Ken Thompson asked me to respond to the following: > I am writing you as you are suppossed to be one of the authors > of the longest known moremover. What's the matter? >[...] > I checked it and yes, it seems to be the record. The question is the > following: is it true that the longest (main) variation has no duals? No. Almost certainly it is not even true that there exists a single longest variation which is not dual-free; this can probably be checked reasonably quickly using Ken Thompson's online database query server starting from http://cm.bell-labs.com/cm/cs/who/ken/chesseg.html, which for the endgame of KRN/KNN will tell the length-to-mate of every winnable position. The record claim was for the largest M for which one can give a legal position without promoted force in which, with best play from both sides, White can mate in M moves but no fewer. No claim is made for uniqueness of any of those moves. Since Ken Thompson's position(*) uses only six men it must be possible to add introductory play to increase Thompson's M=262 by a few more moves. Early this year I gave the following example with M=265: Kc7,Rb7,Rd3,Nc1,Nb8/Kg2,Qg1,Nf2,Nf6: mate in 265 starting with 1 Rg3+ K:g3 2 Ne2+ Kg2(h2) 3 N:g1 K:g1 4 Kd6 etc. [N="S"=Knight; Black alternatives surely lose in at most 200!] This is as far as I know the current record though I'm sure one can add a few more moves some other way. Sincerely, --Noam D. Elkies (*) Actually the position was found by Lewis Stiller, but its length-to-mate was only computed several years later by Thompson, and/or Eugene Nalimov who according to Ken "solved the 262 mover about the same time".
> Another question - where was your problem published for the first time? > Somewhere in > print, or in mailing list or on some www page? This is important as the primary > source > is usually published with diagram as the names of authors are too. the first publish was on my web site http://cm.bell-labs.com/who/ken/chesseg.html the sub-page with diagram is http://plan9.bell-labs.com/magic/eg/wkc7bkg1wrb7wnb8bnf6bnf2m it was finished and published in early jan 2000. the other "author", peter karrer, is completely unknown to me. unless he has a verifiable independent claim, i suspect he got the moves from my web site and added his name. as i understand the sequence of events, another computer endgame specialist, Eugene Nalimov, got the same position a few days after me. i would not mind if he is credited completely or in part. at the least, you should not accept so much inscrutable computer analysis without independent verification. i think nalimov can probably supply it. you have already heard from noam elkies, who prepended some play to extend the record. ken
a) Mr. Neweklowsky forgets to mention his problem is purely based on analysis done by Lewis Stiller. Neweklowsky's contribution is only the first 3 moves. b) It's unknown whether black has a defense that's gonna take white longer to mate. Stiller only searched for defenses which would postpone material conversions (i.e. captures) as long as possible. So it is not unthinkable that black could sacrifice one of his knights somewhere to postpone the mate! I don't feel like checking this though. You could ask someone with access to the 5-piece database. But see c). c) Peter Karrer found a mate in 262, checking Stiller's 6-piece endgame tablebases for distance-to-mate, instead of distance-to-conversion, breaking Mr. Neweklowsky's claim. It is possible that LN's problem actually -is- a longer mate (see b)) though. Included is a PGN file with the Karrer/Stiller game. Joost [Event "Mate in 262!"] [Round "-"] [White "Peter Karrer & Lewis Stiller"] [Black "-"] [Result "1-0"] [FEN "1N6/1RK5/5n2/8/8/8/5n2/6k1 w - - 0 1"] [SetUp "1"]
Dear Juraj: I just looked at the new "longest moremover without obtrusive pieces." No, I didn't cook it. But it DOES have an obtrusive White Bishop! I don't get it... Best, Gianni
Comments to Juraj Lörinc.
Back to main page of Chess Composition Microweb.