Just throwing an idea out here, but would it be possible to have two separate ladders for CAP? One that has the 4th gen CAPmon included, and one that doesn't? That would seem the most judicious answer to the problem. On the other hand, if that isn't possible or would split resources (ie. playerbase) too much, I would be inclined to lean more toward allowing our current CAPs in 5th gen CAP, for a number of reasons. Firstly because I'd prefer not to 'archive' the sum result of several years' experimentation, but also because I'm intrigued to learn about how our creations will cope in a completely new metagame.
The biggest problem, as far as I can see, with deciding this issue now is that we don't actually know anything about the 5th generation that is relevant to changes in existing Pokemon. Certainly there may be new abilities, moves, items, but we can't really say yet that we know how this will be structured. At the moment, the CAP Process contains Concept, Type, Ability, Stats, Movepool, Art and Sprites. Obviously, the concept won't be changed, the type and stats won't be changed unless there is something ingame whereby they change (again, needs looking into). The art is irrelevant, and although the sprites will need changing:
...that is really a secondary concern. As for the abilities, most CAPs already have two abilities anyway, and the only one who doesn't is Arghonaut, who isn't likely to get a new ability - he's got a fantastic one already, he doesn't need another. Barring something like Dream World or 3 abilities for all Pokemon, this can be disregarded. So, the only thing worth bothering about is the movepool.
Obviously, we will have to look at the precedent ingame before making any standing assumptions, but if we take the 4th generation as precedent, the most notable traits standing out are that all moves currently in the movepool remain, as regards TMs, Levelup and Egg moves. What would be very interesting is if we could establish a relationship between the movepool size and moves gained in the shift between generations for the gamefreak-created Pokemon, and the equivalent for VGMs (which would also have to be changed), and then we could either add moves by community consensus as appropriate, or simply reopen movepool submission stages in accordance with the new requirements, much like what was recently done with the current requirements.
One thing I would prefer not to see is any attempt to 'tweak' or 'revise' the CAPs to make them fit in to the 5th gen metagame, even if they end up catastrophically bad or Uber after our first attempt at 'migrating'. They were created at a specific time and should remain largely as such - this is not saying that all new moves should be purely flavour, but we should not be trying to deliberately integrate them. Like I said, I would be greatly interested to see how they perform in a new metagame. So, ideally I would like to see a process of 'make changes' ---> 'test' ---> 'leave alone' (exempting, of course, any revisions or changes to requirements that make updating necessary), where 'make changes' consists of compiling a list of mechanics changes that should be taken into consideration (ie. new abilities, moves being nerfed, type-move reqs. etc)working out new VGMs and VGM/Move limits, revising movepools one at a time, and then playtesting.
I think that's all I want to say.