In the CODEX for Chess Composition Rulebook, you always-always assume castling is okay....unless you're told otherwise.
http://www.saunalahti.fi/~stniekat/pccc/codex.htm
See....Part One / Chapter IV / Article 16 / (1).
Not true; the article referred to by Remellion takes precedence over this one (see PRA). The Codex and FIDE lawbooks, like most lawbooks, must be interpreted in their entirety.
It seems to me the OP's problem should be in the running for, not the hardest, but the most peculiar mate in 4 of all time.
It is because the castling rights play a role, the position is otherwise simple 
There are plenty of problems where castling rights play a role, but how many have you seen where the solution requires arbitration between clashing castling rights? This one is a first for me.
There are plenty of problems where castling rights play a role, but how many have you seen where the solution requires arbitration between clashing castling rights? This one is a first for me.
There are quite a lot though and better than the current one which is a bit fat and scarce on retro-analysis. And they are the single reason that article 16.3 is as it is. It is not (as Remellion assumed) to create the extra challenge of dicovering the RS/PRA nature of a composition but to get rid of distinctive instructions in the problem stipulations. Retro Strategy is mostly just the monkey trick of 'mutually exclusive castlings' demonstrated by this problem plus a bit of uncontroversial 50M stuff.
Nevertheless, the generic subject of Retro Strategy exists and has many potential applications, particularly in fairy chess forms. I entered an orthodox one in Gserpers article on 'most difficult problems' which is much more deservant of that title than the one in this thread. I will repeat it in a later post.
The major issue with these extensions is that they go under the heading 'conventions' and not 'rules'. Conventions relate to 'what is usually done' and 'what is agreed upon' but they are not mandatory in any way. [...]
Basically, it works the other way. The problem-author and the solvers recognize that a problem is of the retro-type. That places us in Rome and invites us to do as the romans - which we do by applying the retro-conventions. Still, they are only conventions; just add 'RS' or 'PRA' to the stipulation to override them.
In my view, conventions for publication and tourneys are, for all intents and purposes, rules. That said, I don't think PRA/RS and 50-move conventions come into conflict with orthodox compositions (except tablebase studies), so there is no need to isolate retro conventions from the rest. As you mentioned before, it is difficult to accidentally (or even intentionally) produce a position with such properties.
Besides, if we were to define what qualifies as a "retro" or "non-retro" composition to determine which conventions apply, analysis of the author's intention is required. It's much cleaner to suppose the retro conventions are always in effect (unless as you said, overridden by the stipulation.)
The primary purpose of the PRA/RS conventions is to ensure there is one universal standard for this otherwise arbitrary business of move rights. And the standard itself is arbitrary - a while back there was this "RV" floating around. I agree that the intention is not to introduce challenge to a problem, but it's nice to have elegant stipulations like "#1" hiding the hardest problems - we couldn't have this if PRA/RV/side to move were required to be stipulated.
@n9531l: Welcome to the world of retros. Enjoy your stay.
As I am sure everybody has realized by now, the solution is dependant on whether or not black has castling rights.
There are many possible games through which the diagram position can be reached, however, from retro-analysis we can at last figure out the following:
_If W reaches the position still having rights to castle, then black will have necessarily lost his right to castle (this can be deduced from retro-analysis).
Some people mentioned that W could just mate B in 2 moves (b7 and b8=Q) if B didn't have the right to castle.
That is exactly the point! If we add the move 1 b7 to the position of the diagram, we could still be following a game in which B had the right to castle and W didn't, and there would still be a scenario where B could just reply to 1 b7 by merely castling.
The only way for W to force a checkmate here is by proving with his moves to be following one of the possible games (that led to that diagram) in which B couldn't castle.
The only way for W to do that is by castling, as that diagram position cannot possibly be reached with both sides still having the right to castle.
The solution, as such will be the following:
1_Qc4 +!! bxc4
Preventing B from castling by giving check and freeing the f1 square to allow the king to castle
2_ 0-0
And now it would be simply impossible to reach the exact same position (including these 2 moves and having also passed from the position in the diagram) with B having castling rights through a normal game.
b7 and promotion of the pawn will follow, and as such we have reached a forced mate in 4 :)
Might be a repetition of what others posted before, still thought I'd still post it as I am the OP. Congrats to those that got it right :D
In my view, conventions for publication and tourneys are, for all intents and purposes, rules. .....
Remellion, I respect your view and I wouldn't expect it to be different. You represent the mainstream convictions in the retrograde world where I am the maverick. Nevertheless I can prove the points I made plus many more but this is not the place to elaborate on them. Just to sow a few question marks in the minds of the traditionalists, I hereby present a composition which effectively devastates the current retro-conventions. It does so by demonstrating they are incomplete and incapable of handling the true generic nature of RS- and PRA-logic (and by implication AP-logic and all fairy forms). But this is only the first page of a 2000 page book awaiting its delivery.
Note: this is just a standard, orthodox, direct mate problem.
Mate in 7 moves
Still waiting for OP's solution...
Still waiting for OP's solution...
It's in #72. I apologize for flooding it.
#73, is a M3 with 1.Bxg6. why are you wanting to play longer?
#73, is a M3 with 1.Bxg6. why are you wanting to play longer?
1.. fxg6 2.Qg8 (or Qf1) 0-0-0 3.Qxf8 (or Rxf8) Bxg5 and now?
#73, is a M3 with 1.Bxg6. why are you wanting to play longer?
1.. fxg6 2.Qg8 (or Qf1) 0-0-0 3.Qxf8 Bxg5 and now?
sorry overlooked the castling 
How do the first three moves of the solution prevent black from castling?
Edit: Oh, I think I get it now. 3-fold repetition forces black's last move to be with the king or the rook.
In trying to solve #73, am I allowed to assume that the position is required to be a legal position? If so, as part of the solution, is it required to explain why Black has lost the right to castle?
Mate in 7 moves
That's another clever puzzle right here
In trying to solve #73, am I allowed to assume that the position is required to be a legal position? If so, as part of the solution, is it required to explain why Black has lost the right to castle?
Though #73 is an embarassment to the designers of the current retrograde conventions, it does not violate any of the real (retrograde) principles of chess compositions. Therefore:
1. The diagram is legal, i.e. there is at least 1 legal proof game for it
2. The solution is legal, i.e. there is at least 1 legal proof game for the diagram + the solution.
3. The solution consists of moves only, it is not coffeehouse chat. To understand it though may require a college course.
Due to the way the Codex lays out the rules, it is in fact necessary to assume that every directmate problem might have a retro component. The PRA/RS conventions (a more detailed analysis given here) apply to all problems - see here examples of confusion when castling conventions (the simple ones, not PRA/RS) they slip into a mainstream publication.
The major issue with these extensions is that they go under the heading 'conventions' and not 'rules'. Conventions relate to 'what is usually done' and 'what is agreed upon' but they are not mandatory in any way. When you wonder about where these things are 'usually done' you come up with one area: the retrograde field. The endgame community will shrug its shoulders about strange conventions dreamt up by retro-weirdos. Not that they won't accept the basic ones - we all need to know that we can or cannot play 0-0 or e.p. - but beyond that the conventions are laid aside at will. This is a good thing since these convention are partly absurd, partly in conflict with FIDE-laws and partly missing where they are absolutely necessary.
Basically, it works the other way. The problem-author and the solvers recognize that a problem is of the retro-type. That places them in Rome and invites them to do as the romans - apply the retro-conventions. Still, these are only conventions; just add 'RS' or 'PRA' to the stipulation to override them.