THE CHAIN OF PROOF IS BROKEN
Suppose a position has a unique proof game for a given number of moves. Does it follow that the prior position also supports a unique proof game? The surprising answer is no, and here are some examples to prove it.
| {A} A.G.Buchanan Original |
{B} Thierry le Gleuher
Hon. Mention, Messigny 98 M6 Phénix 66, September 98 |
{C} A.G.Buchanan 3943 Phénix 106, April 2002 |
||
 |
 |
 |
||
| (15+16) a) PG in 5.0. b) Is the position before Black's last move a correct PG in 4.5? | (16+15) a) PG in 4.5. b) Is the position before White's last move a correct PG in 4.0? | (15+13) a) PG in 6.0 moves. b) Is the position before Black's last move a correct PG in 5.5 moves? |
OK, so you can see that it's just a trick involving castling or en passant. Michel Caillaud's example shows how the trick can be harnessed to ensure the integrity of a deep position.
My ambitions are more humble. In {E}, {F} & {G}, I am looking for quick e.p. positions where there is no check in the final position. {H} is I think the shortest PG of this type involving castling. Compare this with Thierry's {B} which is the shortest PG involving e.p.
| {D} Michel Caillaud Probleemblad, 1998 |
{E} A.G.Buchanan 3944 Phénix 106, April 2002 |
{F} A.G.Buchanan Original |
||
 |
 |
 |
||
| (13+12) a) PG in 22.5. b) Is the position before White's last move a correct PG in 22.0? | (15+15) a) PG in 6.0 moves b) Is the position 2 moves ago a correct PG in 5.0 moves? | (13+16) a) PG in 5.5. b) Is the position 2 moves ago a correct PG in 4.5? |
| {G} Thierry le Gleuher
version A.G.Buchanan Original |
{H} A.G.Buchanan 4304 Phénix 120, August 2003 |
|||
 |
 |
 |
||
| (15+15) a) PG in 5.0 moves b) Is the position 2 moves ago a correct PG in 4.0? | (15+14) PG in 5.0. Is the position before Black's last move: a correct PG in 4.5? |