For each piece, which is the shortest proof game without duals, in which this piece mates? Some additional conditions may be considered, e.g. for a original piece, for a promoted piece, etc.
Many new records in this genre originate in mail discussions about "shortest mates in proof games" on the Retro Mailing List in May 1996 and October 2002. Some of the compositions have been improved in the meantime, but they are nevertheless interesting originals.
Many thanks to Norbert Geissler, who collected and documented the 1996 results, and Christoph Fieberg, who compiled the current list.
The compositions are ranked according to a) less number of moves, b) less number of capturing moves, and c) publication date. If you are aware of better results, or earlier results with the same number of moves, please contribute!
Note: The systematic of this overview and Norbert's work is quite different. Currently, Norbert's work covers also other types of records.
The problems are "numbered" as follows:
Mxy Type 1: Mate by piece xy PxyZ Type 2: Mate by Pawn xy promoted to Z DxyZ Type 3: Mate by piece xy discovered by piece mn Exy Type 4: Mate by Pawn xy capturing e.p. Cxymn Type 4: Mate by 0-0 or 0-0-0, xy = King, mn = Rook
[-nr] appended to one of the "numbers" above: A problem which is an original, but not a record (any more).
xy and mn are the coordinates of the pieces in the starting position.