Different algorithms will usually get different false positives. This is why large collections of algorithms generally perform much better than the best individual algorithm. I imagine these are all weird edge cases as it is.