The difference between approximate and exact false positive rate calculations in real-world scenarios is insignificant. It's fine to use the approximation. If

The difference between approximate and exact false positive rate calculations in real-world scenarios is insignificant. It's fine to use the approximation. If you had very specific tolerances here you wouldn't work in terms of the expected false positive rate anyway; you'd build your filter and validate on the real thing.