So... since we have learned that since the probability terms in the CH inequality can only range from 0 to 1/2 because it only deals with + and ++ counts, then I believe the CHSH probability version based on CH would be the following.
|p(a, b) - p(a, b') + p(a', b) + p(a', b')| ≤ 1
Remembering that we are only dealing with ++ counts for each term. So we could have for independent terms,
1/2 - 0 + 1/2 + 1/2 = 1.5
So 1.5 is the absolute bound for independent terms of this version. Another thing is that an experimenter would have to know N the total number of emissions for each term. So this version is not reallly practical as far as experiments go.