I inserted M items into a hash table with N buckets, and found that the largest bucket has K items. How can I compute the p-value corresponding to the null hypothesis that my hash function is uniform?
I know that one can derive various with-high-probability asymptotic bounds on K as a function of M and N. But it's not obvious how to utilize those asymptotic bounds to get a concrete p-value.
My practical use-case is that I want my application to automatically report a warning if it detects poor hash-balancing. Something this theoretically grounded is probably overkill, but why not use something sound if possible.
