2 ms·
That just explains Bayesian inference in general, which is useful, but not what I'm after. I was specifically interested in the application to binary search /
by eru 2y ago
That just explains Bayesian inference in general, which is useful, but not what I'm after.
I was specifically interested in the application to binary search / bisection in the presence of flaky tests.
- dwattttt 2y agoYou apply the same process, but your hypotheses are "bug is before/after this bisection point". Your "probability of evidence given hypothesis before/after" are where you incorporate your guess about the tests flakiness. Still works even if you don't have "true" numbers for the tests flakiness, just won't converge as quickly