I think the point has to be on how much you expect the game to change over the next few months. With LambdaRogue I'm guessing there's a lot of scope for the new crawler mode to actually change quite a bit. Plus feedback can inspire significant alternations. But if it doesn't work out like this then I guess it may not be suitable for it to stay in.
There could be a separate strand for more mature roguelikes to polish up, add features, etc. But I'd say it's best to start inclusive and split into sub-groups later if we feel it appropriate. The balance should definitely be more towards new games, but developing established games is certainly not to be discouraged.