# Reverse Puzzling - Curious Chess

George rubbed his hands together. "What have you got for me today?" he asked.

I showed him the pieces set up on the chess board and told him the stipulation.

"That's a very strange one." He said. "Where did it come from?"

I told him it was written in 1874 and the name of the author.

"Interesting, interesting. So you can't move the pawns, eh? Mmm. Well, Dijkstra should do it no problem."

My eyes widened at the thought but George was nothing if not thorough.

Half an hour later, I returned and George proudly announced the answer gesturing grandly at the whiteboard: "Twenty-seven! And there are two ways to do it."

I looked in awe at the whiteboard.

What was the puzzle that I gave George?

Well, this is frustrating. The puzzle is clearly

the one described under the heading "logjam" here but the actual problem is given as a 4x4 table of images, and the images no longer exist, and my attempts to find the problem elsewhere on the internet have failed. But I think the position is

. P P P
. P R P
P P B K
n . B R

and the stipulation is that only W is to move, and that he is to capture the BN on what I would call a1 but I suppose we should in fact regard as e1 (so that this is the "southeast" quadrant of an ordinary 8x8 board), without moving any of the pawns.

The name of the author is

W A Shinkman, who I think was quite a prolific composer of ingenious chess problems. The citation on the page I linked above says that this problem is from the Deutsche Schachzeitung in June 1874.

Incidentally, my response to the problem would be pretty much the same as George's, though I'd be more inclined to get a computer to solve it rather than trying to do it on a whiteboard.

I believe that this is a:

Move tree map of a chess maze (I couldn't find it online but here is a rendition, click to take to a board on lichess):

with the rules that:

No pawn may be moved and the objective is for the white king to take the black knight.

George is trying to:

Find the minimum number of moves to win and find how many ways there are to do it

and has used:

Djikstra's algorithm to navigate the move tree with the stipulation that no position may be repeated (so that the graph isn't infinite)

• Less than an hour! Boom! Nice! Was it too easy, or were you just that good? Commented Feb 3, 2019 at 3:28
• I sort of got what was going on from the start. Unfortunately Gareth managed to ninja me. Commented Feb 3, 2019 at 23:46
• He does that a lot... Commented Feb 4, 2019 at 13:40