[Retros] the risk of fractional moves

raosorio at fibertel.com.ar raosorio at fibertel.com.ar
Sun Mar 2 10:20:51 EST 2008

Fractional moves without full retrodemonstration are very risky indeed.

I use here the Gligor's and Per's positions as examples.

Ke1-g8 is half a move but to remove an opponet's piece from the board
without moving the capturing piece is half a move too.

So, in Gligor's position there are many variants like Bc1-f4 (completing a
knight capture that was already removed from f4) and Nh4-f5 (the
same). Both would be #0,5, as legal as the half castling variant.

In addition, to install a wQ on b8 is also legal as a fractional move , completing
the 4th part of the move c7xRb8 (the 3 first parts already done: to remove the
bR, to reach the 8th rank with the pawn and to remove the pawn). this would
be #0,25.

All these variants solve the "last black move" issue.
Apparently, the problem should be presented as "mate in exactly 1.5 single moves"

the same applies to Per's twin a). Say, g6xQf5 is on the process, so black retracts
the first part of the capture (reinstalling the wQ on f5) and then 1.Kf8, Qxf7#

Hi, Roberto

See that problem

Gligor Denkovski & Miroslav Stosic
Mlad borec 1971 (v)

Kg1 Rh1 Bc1 Bh5 Sg6 Sh4 Ph3 - Kg3
1.Bf4#? 1.Sf5#? No last black move!
Retro: The last black move was Kg2-g3, before wK move from e1 to f1 (first
part of 0-0).
1.Rf1-f1!! (ends 0-0) Kxh3 2.Rf3#


Hi all!

Note that there has been a discussion concerning incomplete moves ('partial/half halfmoves'). Here one of my problems (PDB 1004346) :

Per Olin, 2nd Hon. ment. Suomen Tehtavaniekat 1994 Kc7 Df4 pawn g5 - Kg8 Th8 pawns e5, f7, g6,h6 (3+6) -1/2 bl, then h#1 b) turn 90 degrees c) turn 180 degrees

a) Retract: 1. Ke8 (0-0), then 1. Tf8 Dxe5# b) R: 1. g2xSf1, then 1. Kh3 Dxg3# c) R: 1. a4xb3 (e.p.), then 1. Ka2 Dxc2#.

Michael Caillaud has made a very entertaining proofgame with two solutions in x.75 moves (or was it x.25); can't find it in the PDB.

Best regards


More information about the Retros mailing list