[Retros] On cylinder castling

[Etienne Dupuis]

Greetings All !

The recent debate about cylindrical castling was quite entertaining. Let me 
jump in. First, the problem I published can be found on 
http://lestourtereaux.free.fr/chess/14.html . When I composed it, as Christian 
said, I never thought that ‘extended castling’ was possible. Why ? There are at 
least two ways to define castling :
1.	Explicitely, by stating that the king on e1 moves to g1/c1 and that the 
rook on h1/a1 moves to f1/c1 if some conditions are satisfied.
2.	Implicitely, by stating that the king moves two squares toward a rook 
of the same color that hasn’t move from its initial square and that the rook 
jumps over the king, again if some conditions are satisfied.

For definition (1), extended castling on vertical cylinder is certainly not 
natural, while for definition (2) it is. It seems that many composers consider 
definition (2) to be valid for vertical cylinder chess.

Let us now consider horizontal cylinder boards. By definition (1), the are two 
kinds of castling. However, definition (2) poses problem; since the initial 
position is illegal, we could consider there is no such initial position and 
that castling is impossible. I am pretty sure most composers will accept 
castling in such a problem.

There are other fairy conditions which may lead to a debate. What about the 
condition that a piece may move and check only if observed by another piece ? 
White king e1, white rook h1 and black knight h2. May white castle ? Probably 
no. Perhaps yes. Also, if there is ‘n’ white pawns and more than ‘8-n’ white 
fairy pieces in a direct mate, do we care that black has a last move since he 
initial position is undefined ? Finally, can a paralysed rook be involved in 
castling since it is usually considered as a king move ?

We get to the crucial point. There is an official set of rules for chess. There 
is none for fairy chess; there is only a tradition. It is perfectly normal in 
that case that there are some ideological debates.

If we come back to my published problem, the editor of StrateGems was in my 
opinion awkward in dealing with the polemic. My problem can be ‘corrected’ 
(solutions with extended castling can be prevented) by changing slightly the 
position of a single piece – see my web site. The problem published as a 
correction is not the correction, it is a variant on the same theme which I 
found funny; using the extended castling the black king gets trapped in the h8 
corner. This confusion is certainly unfortunate. As Christian, I do not 
consider my original problem to be really cooked, however since it is extremely 
easy to please everyone by moving one piece that I see no reason to not do it


Cheers,

Étienne


P.S. When trying to define mathematically the rules of (fairy) chess I ran into 
big problems. The board is defined without problems. But then a piece seems to 
be defined by it’s moving abilities, which depend on the position, composed of 
pieces on the board. I tried to circumvent the problem by ‘naming’ pieces, but 
this is possible only if the number of different pieces is countable, which I 
think is not since it is basically an automorphism from the space of positions 
to itself. I put everything aside thinking that I must be either mad or 
insufficiently bright !