# [Retros] Shortest SPGs ending in mate

Francois Labelle flab at EECS.Berkeley.EDU
Mon Jul 12 18:27:34 EDT 2004

It looks like in 1996 a lot of you had fun on the mailing list trying to
beat the records for the shortest SPGs ending in specified types of mates.
Now with a computer I was able to verify a lot of these SPGs and beat a
few records.

I generated every checkmate proof game up to 5.0 moves, and every
one-sided checkmate proof game up to 11 moves. (By elimination, this
proves optimal the checkmate proof games in 5.5 moves and one-sided ones
in 12 moves that I couldn't beat.)

Below are the problems for which I beat the record (I write the previous
record, followed by my improvement). When there is a choice I give the
first example in alphabetical order of FEN. I also mention the problems
for which there is exactly one example of the minimal length.

Then the same thing is repeated for one-sided chess.

Thanks to Richard Stanley for sending me a summary of the 1996 discussion.

Notation:
rc : mating move Requires Capturing
wc : Without Capturing in the mating move
mc : no Capturing during all the moves of the Mating side
nc : No Capturing at all
dc : mating move is a Double Check

====================================
Improvements:

Mate by knight (3.0):
rc: G"oran Wicklund, retros mailing list, october 1996
(5.5): 1.Nf3 e6 2.Ne5 Ne7 3.Nxd7 e5 4.Nxf8 Bd7 5.Ne6 Rf8 6.Nxg7#
Francois Labelle
(4.5): 1.d4 e6 2.Bg5 Ba3 3.Nxa3 Qe7 4.Nb5 Qf8 5.Nxc7#

Mate by pawn promoting to queen (4.5):
wc: Norbert Geissler, retros mailing list, may 1996
(5.0): 1.d4 b5 2.Qd3 b4 3.Qa3 b3 4.Bd2 bxc2 5.Bc3 c1Q#
Francois Labelle
(4.5): 1.c4 d5 2.c5 Qd6 3.cxd6 Bd7 4.dxc7 Bc6 5.c8=Q#

Mate by double check, no checking piece is attacked (minimal length ?):
wc: open
Francois Labelle
(5.0): 1.e3 g6 2.Qg4 Bg7 3.Qxd7+ Bxd7 4.Ke2 f5 5.Kd3 Bb5#

Mate by double check, both checking pieces are attacked (minimal length
?):
rc: open
Francois Labelle
(5.0): 1.d4 Nf6 2.Bf4 Nd5 3.Bxc7 Nb4 4.Bxb8 Qa5 5.Bc7 Nxc2#
mc: Unto Heinonen, retros mailing list, october 1996
(6.0): 1.Nf3 Nc6 2.Ne5 Nd4 3.Nc4 e5 4.f4 Qf6 5.Kf2 Qb6 6.Kg1 Nf3#
Francois Labelle
(5.0): 1.d4 Nf6 2.Bf4 Nd5 3.Bxc7 Nb4 4.Bxb8 Qa5 5.Bc7 Nd3#

------------------------------------
Those were the unique examples of the stated length:
(so asking to find the proof game is a sound problem)

Mate by queen (2.0):
wc: Juha M. Saukkola, retros mailing list, june 1996
(2.5): 1.e4 f5 2.exf5 g5 3.Qh5#

Mate by bishop (2.5):
mc: Richard Stanley, retros mailing list, june 1996
(3.5): 1.e3 h5 2.Bd3 h4 3.Bh7 f5 4.Bg6#
nc: Richard Stanley, retros mailing list, june 1996
(3.5): 1.e3 h5 2.Bd3 h4 3.Bh7 f5 4.Bg6#

Mate by double check, one checking piece is attacked (4.5):
nc: Richard Stanley, retros mailing list, october 1996
(5.0): 1.f4 h5 2.Kf2 Rh6 3.Kg3 Rf6 4.Kh4 e6 5.Kg5 Rf5#

====================================
Improvements in the one-sided case:

Mate by rook:
dc: Richard Stanley, retros mailing list, june 1996
(10): 1.a4 2.Ra3 3.Rd3 4.Rxd7 5.d3 6.Bg5 7.e3 8.Qf3 9.Qc6 10.Rxe7#
Francois Labelle
( 9): 1.h4 2.Rh3 3.Rd3 4.Rxd7 5.d4 6.Bg5 7.e3 8.Bb5 9.Rxe7#

Mate by bishop:
wc: Juha M. Saukkola & Gianni Donati, retros mailing list, october 1996
(10): 1.a4 2.Ra3 3.Rf3 4.Rxf7 5.Rf3 6.Rc3 7.e3 8.Bd3 9.Bxh7 10.Bg6#
Francois Labelle
( 9): 1.h4 2.Rh3 3.Rc3 4.d3 5.Bf4 6.Bxc7 7.Bxd8 8.Rxc8 9.Bc7#

Mate by knight:
dc: open
Francois Labelle
(11): 1.c4 2.c5 3.c6 4.cxb7 5.bxa8=N 6.Nb6 7.Nxd7 8.Nxb8 9.Nd7 10.Qa4
11.Nf6#

Mate by promoted pawn (knight):
dc: open
Francois Labelle
(11): 1.c4 2.c5 3.c6 4.cxb7 5.bxa8=N 6.Nb6 7.Nxd7 8.Nxb8 9.Nd7 10.Qa4
11.Nf6#

Mate by a battery without double_check:
rc: Gianni Donati, retros mailing list, june 1996
(13): 1.e4 2.Be2 3.Bh5 4.Ke2 5.Kf3 6.Kg4 7.Nf3 8.Re1 9.Re3 10.Ne1
11.Rf3 12.Rxf7 13.Rxg7#
Francois Labelle
(11): 1.e4 2.Be2 3.Bh5 4.Nf3 5.O-O 6.Re1 7.Re3 8.Ne1 9.Rf3 10.Rxf7
11.Rxg7#
wc: Richard Stanley, retros mailing list, june 1996
(10): 1.a4 2.Ra3 3.Rg3 4.b3 5.Bb2 6.Bxg7 7.Bxf8 8.Rxg8 9.Rxh8 10.Bg7#
Francois Labelle
( 9): 1.h4 2.Rh3 3.Rc3 4.d3 5.Bf4 6.Bxc7 7.Bxd8 8.Rxc8 9.Bc7#

------------------------------------
Those were the unique examples of the stated length:
(so asking to find the proof game is a sound problem)

Mate by bishop:
wc: Francois Labelle
( 9): 1.h4 2.Rh3 3.Rc3 4.d3 5.Bf4 6.Bxc7 7.Bxd8 8.Rxc8 9.Bc7#

Mate by knight:
wc: Timothy Luffingham, retros mailing list, june 1996
( 9): 1.Nc3 2.Nd5 3.c3 4.Qc2 5.Qxh7 6.Qxg8 7.Qxg7 8.Qe5 9.Nf6#

Mate by promoted pawn (queen):
rc: Timothy Luffingham & Gianni Donati, retros mailing list, june 1996
( 6): 1.d4 2.d5 3.d6 4.dxe7 5.Qd6 6.exf8Q#

Mate by promoted pawn (rook):
rc: Timothy Luffingham, Gianni Donati, retros mailing list, june 1996
( 6): 1.d4 2.d5 3.d6 4.dxe7 5.Qd6 6.exf8R#

Mate by promoted pawn (bishop):
rc: Michel Caillaud, retros mailing list, october 2002
(11): 1.e4 2.Qh5 3.Qxh7 4.Qh3 5.Qb3 6.h4 7.h5 8.h6 9.h7 10.hxg8B
11.Bxf7#

Mate by a battery without double_check:
rc: Francois Labelle
(11): 1.e4 2.Be2 3.Bh5 4.Nf3 5.O-O 6.Re1 7.Re3 8.Ne1 9.Rf3 10.Rxf7
11.Rxg7#
wc: Francois Labelle
( 9): 1.h4 2.Rh3 3.Rc3 4.d3 5.Bf4 6.Bxc7 7.Bxd8 8.Rxc8 9.Bc7#

------------------------------------

Francois