Solutions for Belfort Proof Games

knight Belfort

.Dupuis, v A.G.Buchanan

Problemesis 20, August 2002
(16+16) PG in 12.5

1.Sf3 Sc6 2.Sd4 Se5 3.Sc6 Sf6 4.d4 Se4 5.Be3 Sd2 6.Sc3 Sb1 7.Qd2 Sf3 8.Kd1 e5 9.Sd5 Bd6 10.Se7 Sg1 11.Sg8 Qe7 12.Sb8. (C+ Natch 2.0)

The published version had the bishops on d6/e3 on a3/h6. The change is purely aesthetic: I like the stepping-stone effect.

For comparison, here is an earlier composition of the same length by Étienne Dupuis, modified slightly to remove the unnecessary capture. These are the quickest positions known today for the capture-free knight Belfort.

bishop Belfort

Frank Christiaans &
Henk Boumeester

Probleemblad, Nov 1993
(16+14) PG in 7.5

1.d4 e5 2.Bd2 Be7 3.Bb4 Bg5 4.Bf8 Bc1 5.e3 d6 6.Ba6 Bh3 7.g4 b5 8.Bc8 Bf1. (C+ Natch 2.0)

Joost de Heer has pointed out a definite partial anticipation, see left. This solves as 1.e4 d5 2.Be2 Bd7 3.Bg4 Bb5 4.Bc8 Bf1 5.d3 g5 6.B×g5 Bh6 7.B×e7 Bc1 8.Bf8. (C+ Natch 2.0)

This is quicker, but not capture-free. Mine has a symmetric position, but it emerges from symmetric play so is not as attractive as {A}.

I would add bPg6 to the Christiaans/Boumeester original,  to remove an inessential capture. I think that 7.5 is as fast as the bishop Belfort can get.

1.d4 f5 2.Kd2 Kf7 3.Ke3 Ke6 4.Kf4 Kd5 5.Kg5 h6 6.Kg6 e5 7.Kf7 Qh4 8.Bg5 Ke4 9.f4 Qe1 10.g3 Ke3 11.Bg2 Kf2 12.Bc6 e4 13.d5 e3 14.Qd4 Qd1 15.Qf6 Ke1 16.Qd8 Be7 17.Ke8. (C+ Natch 2.0)

This is the quickest known, and is also capture-free.

rook Belfort

Thierry le Gleuher

Messigny, 1996, version
(11+12) PG in 12.0

1.f4 a5 2.f5 a4 3.f6 a3 4.f×g7 a×b2 5.g×h8=R b×a1=R 6.Bb2 Bg7 7.Bc3 B×c3 8.d×c3 h5 9.c4 h4 10.c5 h3 11.c6 h×g2 12.c×b7 g×h1=R 13.b×a8=R. (C+ Natch 2.0)

I am largely anticipated by Thierry le Gleuher (see right), although I am happy to have achieved the awaybase position.

1.g4 b5 2.g5 b4 3.g6 b3 4.gxh7 bxc2 5.hxg8=S cxb1=S 6.Qb3 g5 7.Qxb8 g4 8.Qxc8 Rb8 9.b4 Rb6 10.b5 Rbh6   11.b6 g3 12.b7 gxh2 13.b8=S hxg1=S 14.Qb7 Rxh1 15.Qb3. (C+ Natch 2.0)

The basic Belfort task is achieved in 13.0, after that the objective is to make the Black rook Belfort and to remove the distracting the White queen from the Black camp (it's not Belfort except on d8, when it would be giving check).

knight Belfort

Joost de Heer,
R65 Problemesis 29, August 2002
v Nicolas Dupont,
v A.G.Buchanan
(11+12) PG in 13.0

Prior art for this task (Joost de Heer's commended R65 & Nicolas Dupont's version - see comments on R65 in Problemesis) uses a completely different matrix. I can see how to adapt Nicolas' version to avoid a check in the final position. (See left.) (C+ Natch 2.0)

1.h4 g5 2.h×g5 h6 3.R×h6 a6 4.R×a6 Rh1 5.Rh6 R×a2 6.Rh8 R×b2 7.Ra8 Ra2 8.Bb2 Ra1. (C+ Natch 2.0)

1.Sg1-f3 Sg8-f6 2.Sf3-e5 Sf6-e4 3.Se5×d7 Se4×d2
4.S×b8 S×b1 5.Bf4 h5 6.e3 h4 7.Ke2 h3 8.Kf3 Ph3×g2 9.Ph2-h4 g1=S 10.Kg2 Bg4 11.h5 e6 12.h6 Ke7 13.h×g7 Kf6 14.g8=S Kg7 15.Bg5. (C+ Natch 2.0)

The final position is symmetric, but play is asymmetric after the prologue of 4.0 moves.

1.a4 h5 2.a5 h4 3.a6 h3 4.a×b7 h×g2 5.Ra6 g×f1=B 6.Rd6 e×d6 7.h4 Q×h4 8.b×c8=B Qe7 9.f4 a5 10.f5 a4 11.f6 a3 12.f×e7 a×b2 13.e×f8=B b×c1=B. (C+)

Bishops generally win any race with pawns - as long as the distance is more than two squares, so it's remarkable that the pawn promotion solution is sound, in what is a totally open position.

Back to compositions

Back to home page