[Retros] Series proof games leading to a checkmate in one move.

Francois Labelle flab at wismuth.com
Thu Nov 4 00:36:51 EDT 2010

Hi Nicolas,

Nice challenge... for a computer!

I searched all such problems for N <= 12, and for N <= 9 + rank(wk),
where rank(wk) is the rank of the white king in the final position. I
found the following improvements (I also list ties if they're different

Pa7, 1 fewer capture
1.e3 5.Kb5 6.c4 7.Qa4 8.b3 10.Bxe7 12.Ba5 a6#

Pe7, 1 fewer move (2 different solutions)
1.c4 3.Qxa7 4.Qe3 5.d4 9.Kd5 10.Qe4 12.Be5 e6#
1.g4 4.Bxh7 6.Ba6 7.d3 11.Kg5 12.Bf4 e6#

Pg7, no improvement but different
1.c3 3.Qg4 4.d4 5.Bg5 6.f4 10.Kh5 11.h4 g6#

Ra8, no improvement but different
3.Nb5 4.d4 7.Kb4 8.c3 10.Qxa7 11.Qxb7 12.Qa6 13.Ka5 14.b4 Rxa6#

Because I did an exhaustive search up to N = 12, all the records with N
<= 12 are optimal. In addition, the record for Pd7 must be optimal
because it has 13 moves and no captures.


I realize I haven't posted here in a while! I got some new computer
results in the last 3 years. I know all the proof games in 5.0 moves,
all the checkmate proof games in 6.0 moves, and all the one-sided proof
games in 11 moves. I haven't data-mined any of it yet, except for this

rn1qkbnr/ppp2ppp/8/3p4/8/1b1P4/PP2PPPP/RNBQKBNR PG in 5.0 moves, 3

which is probably the best 3-solution PG of that length, and was used in
the Messigny 2008 retro solving contest.

Also Berkeley just closed my e-mail and web accounts, so I'm moving my
chess page from
http://www.cs.berkeley.edu/~flab/chess/chess.html to my own domain at

My new e-mail address is the one I'm using for this message.


On Fri, 2010-10-22 at 12:37 +0200, Nicolas Dupont wrote:

> Dear retro fans,


> The following series (exact) proof game has been discovered some years

> ago, by various members of the french website France-Echecs :


> 1.c3 2.Qa4 7.Ka5 8.b4


> It leads to a checkmate in one move by the black side, namely c6#, and is

> probably the shortest series proof game permitting such a #1.


> You can regard this as a kind of helpmate series problem, beginning with

> the initial game array (and with reverse colors, i.e. we prefer white to

> begin than black, this is more natural in that setting).


> Here the last move is with the c-pawn, and N, the number of white moves,

> is 8. Andrew Buchanan and I proposed to extend this result, by asking :


> For each of the remaining black 15 pieces, which is the minimum value of N

> achievable ?


> The massed brains of France-Echecs have achieved the following records

> (thanks to Popeye, each of them is C+). Will you be able to beat some of

> them ? Note that the above question makes also sense for black's 0-0 and

> 0-0-0, but we were unbable to find a solution in those tricky cases. Note

> also that the black move has not to be unique, neither giving check, i.e.

> we only need that a move by the selected black piece leads to a checkmate.


> Our purpose is to write an article incorporating the best results, with

> the scope that they are really the length records. That is why we need

> your help... If you are able to show a game with same number of

> moves, but fewer captures, this is considered as an improvement.


> Thanks and all the best,


> Nicolas.


> Pa7 (Nicolas Dupont) 12 moves

> 1.e3 2.Bb5 3.d3 7.Ka5 8.b4 10.Bxg7 12.Bb6 and then axb6#


> Pb7 (Arthur Dupont) 11 moves

> 1.a4 3.Rh3 4.e3 5.Bb5 6.d3 10.Ka5 11.b4 and then b6#


> Pc7 (France-Echecs) 8 moves

> 1.c3 2.Qa4 7.Ka5 8.b4 and then c6#


> Pd7 (Jean-Christian Galli) 13 moves

> 2.a5 4.Re4 5.d4 9.Ke5 10.Bf4 11.e3 13.Bd5 and then d6#


> Pe7 (Jean-Christian Galli) 13 moves

> 2.h5 4.Rd4 5.e4 9.Kd5 10.Bc4 11.d3 13.Be5 and then e6#


> Pf7 (Jacques Dupin) 12 moves

> 1.h4 4.Rxd7 6.Re4 7.d4 11.Ke5 12.Bf4 and then f6#


> Pg7 (Arthur Dupont) 11 moves

> 1.h4 3.Ra3 4.d3 5.Bg5 6.e3 10.Kh5 11.Qg4 and then g6#


> Ph7 (Nicolas Dupont) 12 moves

> 1.d3 2.Bg5 3.e3 7.Kh5 8.g4 10.Bxb7 12.Bg6 and then hxg6#


> Ra8 (Andrew Buchanan) 14 moves

> 4.axb7 5.Rxa7 7.Rh3 8.d3 10.Nf3 11.Bd2 12.Qa1 14.Kc1 and then Rxa1#


> Nb8 (Jacques Dupin) 12 moves

> 1.h4 4.Rxd7 6.Re4 7.d4 11.Ke5 12.Bf4 and then Nc6#


> Bc8 (Jean-Christian Galli) 14 moves

> 4.dxe7 5.exf8=B 7.Bfg3 8.Bcf4 9.e3 13.Kh3 14.Qxd7+ and then Fxd7#


> Qd8 (Andrew Buchanan) 10 moves

> 4.fxe7 5.exf8=N 6.Nxd7 9.Nh6 10.g4 and then Qh4#


> Ke8 (Nicolas Dupont & Jean-Christian Galli) 28 moves

> 1.a4 4.Rxh7 6.Re3 11.hxg8=N 13.Nxf7 14.Nxh8 16.Nxf8 19.Nfb5 20.d4

> 27.Kh8 28.Rh7 and then Kf7#


> Bf8 (Nicolas Dupont & Andrew Buchanan) 15 moves

> 1.a4 4.Rxd7 5.d3 6.Bf4 7.e3 9.Bh5 10.f3 13.Kh4 14.Bg3 15.Rxe7+ and then

> Bxe7#


> Ng8 (Nicolas Dupont) 11 moves

> 1.h4 4.Rg5 5.g3 6.Bh3 11.Kh5 and then Nf6#


> Rh8 (Andrew Buchanan) 12 moves

> 4.hxg7 5.Rxh7 7.Ra3 8.e3 10.Nc3 11.Be2 12.Kf1 and then Rh1#





> _______________________________________________

> Retros mailing list

> Retros at janko.at

> http://www.pairlist.net/mailman/listinfo/retros

More information about the Retros mailing list