Representativeness

I came up with this today.

You could say a theory is x% representative, which you get by dividing the number of cases where it works by the number of all possible cases.

For example, take a function F(a,b) which behaves as a logical OR. The theory that it behaves as a logical AND is then 50% representative: it works for (0,0) and (1,1) (2 cases) and fails for everything else (there being a total of 4 possible cases).

For dealing with infinities it could be tricky.

For instance you could say the statement that all integers greater than 0 are even is 50% representative, because of its periodicity.

But then for the theory that all positive integers are non-prime you would have no periodicity; primes get rarer and rarer and in the end you'd end up with 100% representativeness, which can be misleading.