Rather a broad, all-encompassing thing to state, right?

Well, it’s true. Statistics don’t prove anything. Even the most rigorously undertaken statistical tests can only strongly suggest whatever outcome the hypothesis the tester set out to examine.  Practically every properly conducted statistical test is built around a confidence level.  That’s to say that there is a % chance that the expected outcome may be incorrect, or the obverse of that, what you’re expected to be incorrect turns out otherwise.

I’m not going to go into type 1 or type 2 errors but in science, most confidence levels for tests are set at 95%. So, there is a 5% chance that data collected falls outside of the test’s purview. Five percent is a significant figure, when all things are weighed up. Even if the confidence level was set at 99%, there still a 1% chance of an anomaly. Even in things one would consider dead certainties.

So, while lawyers may be able to prove things, statistics can’t – and it’s not their task to. They can only strongly suggest.