If you could distill your write-up to a formula, that would be helpful. -- I already did that

I guess that your idea of a formula and mine are different.

where $b1 is the number of bits set in the first vector.

The "number of bits set" in each vector at any given insert is entirely dependent upon not just how many, but what values have already been inserted.

Which means that to use your description to calculate the probabilities, I would need to iterate the entire thing and count the number of bits set in each vector at each iteration. And then those "calculated probabilities" would only be applicable to that particular sequence of inserts.

At which point, I might just as well just record the actual numbers of false positives, rather than calculate them.

You'll no doubt come back and tell me that I've completely misunderstood you (again), and that everything I need to know is right there if I would only read you correctly.

Thank you for your attempts to assist me.