﻿﻿ MatPlus.Net

Website founded by
Milan Velimirović
in 2006

20:44 UTC
 ISC 2021

Remember me

 CHESS SOLVINGTournamentsRating lists1-Apr-2021

MatPlus.Net Forum General Proof Games without captures

### Proof Games without captures

Are there any orthodox proof games without any captures? Obviously all 32 pieces would have to be present in the final position.
If yes, what is the length record?

I think the record is 43.0: https://yacpdb.org/#306377 (Thierry le Gleuher has published a 43.5 version but it was cooked, as far as I remember), but it may be cooked.

EDIT: Cooked (https://pdb.dieschwalbe.de/P0002554)

PDB gives almost 200 results via " G='Retro' and APIECES=32 AND STIP='BP in' AND K='Unique Proof Game' "
But I'm not an expert here in this field, so I don't know, which problem is currently the longest sound proof game.

Ps.: Possibly https://pdb.dieschwalbe.de/P1386520 with 42.5 moves.

The longest *symmetrical* one I know is in 30.0 moves: https://pdb.dieschwalbe.de/P0000345

Unto Heinonen, Die Schwalbe 144 12/1993, PG 30.0
(= 16+16 )

Frank wrote:
QUOTE
PDB gives almost 200 results via " G='Retro' and APIECES=32 AND STIP='BP in' AND K='Unique Proof Game' "

Great! This gives 184 answers. A simpler query " K='Unique Proof Game' and apieces=32 " gives 232 results.

Why is this?

(1) The stipulation field is not multi-language yet and there are some stips in English or French.

(2) Worse there many unique proof games without the sub-string 'bp in' the stipulation but e.g. 'Beweispartie', 'bp?' or "kbp?' Guessing sub-strings is a fumbling business. The query " K='Unique Proof Game' and stip='bp' and not STIP='BP in' " gives 84 results.

(3) K='unique p' is a subset of g='retro', so don't need the latter.

(4) BTW: there is also k='capture-free', which is already tagged to all 119 capture-free non-fairy unique proof games. But that is like aristocrat, or miniature: kind of icing on the cake, and should really be added automatically. It's useful for a novice exploring what kind of features are interesting to composers, more than for a experienced person searching for what problems have specified features, for which apieces=32 is fine.

My default querying approach in PDB is to ask first very simple keyword queries, and then refine if necessary. If I get any false positives, I correct the keywords as soon as possible. So every query helps to improve the data quality for those PGs we have. The real PDB quality hole is that we need more recent problems uploaded.

Thanks,
Andrew

There is a longer symmetrical proof game in 34 moves, unfortunately cooked:
https://pdb.dieschwalbe.de/P0005310
If I remember correctly Bachmann showed a correction at "Schwalbe Tagung" in Göttingen.
The cook started at 28th move, so there are good chances to exceed Heinonen's record.

I agree with Frank that the leader right now is Unto with 42.5.

The top five longest capture-free unique proof games as far as I can see are:

42.5 Unto Heinonen, Strategems 2020 (v) https://pdb.dieschwalbe.de/P1386520
38.5 Michel Caillaud, Die Schwalbe 1983 https://pdb.dieschwalbe.de/P0002187
Then a big gap...!
30.5 Kay Soltien & Theo Kuner, Die Schwalbe 1961 https://pdb.dieschwalbe.de/P1388821
30.0 Unto Heinonen, 8332 Die Schwalbe 144 12/1993 https://pdb.dieschwalbe.de/P0000345 the symmetric one!
29.5 Nicolas Dupont & Michel Caillaud, Die Schwalbe 2018 https://pdb.dieschwalbe.de/P1388822

This list comes from the PDB query:
k='unique p' and apieces=32 and (sol='30' or not sol='1') and not g='Fairies' and not cooked

None of these is C+ but none have been cooked. All other candidates I know of (including Unto's 43.0 1st Prize in Probleemblad 1991) have been cooked.

Thanks,
Andrew

Going *non-unique*, the longest is 91.5 moves. https://pdb.dieschwalbe.de/P1374185

Yoav Ben-Zvi, BZ3 Feenschach 01/2014, After Karl Fabel, PG 91.5
(= 16+16 )

Thanks, Everyone!
That's a lot of "Pacific PGs"!

Now, are there similar, possibly longer, ones in the "From position A, reach position B" PG category?
Here, you _could_ have less than 32 pieces and Pawns piled up on files, but obviously in a legal position (though theoretically, illegal positions would be possible and may give longer sequences!).

(10) Posted by Joost de Heer [Thursday, Apr 22, 2021 08:26]

There are some very long ortho-reconstructions, which is a form of A->B chess without captures.

Jean-Michel Trillon
Dedicated to J. Rotenberg
Rex Multiplex 1982, prix spécial
(= 3+58 )

Ortho-reconstruction in 32551.5 moves

The longest dual-free ortho-reconstruction with a 'normal' position I could find is:

Michael Schreckenbach
Die Schwalbe 2015
(= 10+13 )

Ortho-reconstruction in 57.5 moves