I agree with all that. But I’m talking about exact integer values as mentioned in the parent.

I just think this has to be true: count(exact integers that can be represented by a N bit floating point variable) < count(exact integers that can be represented by an N bit int type variable)