Well, the reason it that we have some constraints on our bounds, because we need to ensure that entities in the bounds can move the maximum movement rate and not exit the collision bounds (the outer bounds). If we allowed entities that were not in the bounds to be in the simulation region simply because they were in a chunk that was in bounds, that entity could theoretically leave the collision region and miss a collision it should have had.
We could instead enlarge the outer collision bound by chunk side, so if we found that it was too onerous to do the extra checking, we could do that. But it doesn't seem like a great tradeoff, I should think.
- Casey