Agile A/B testing with Bayesian Statistics and Python

Автор предлагает анализировать результаты A/B-тестов с помощью Байесовских методов, которые во многом лучше используемых повсеместно способов.

Цитируем:

«There are three major advantages the Bayesian A/B test has over the frequentist method. The first is that it’s far easier to interpret the results. For example, you can easily compute the probability that version B is better than version A. The second major philosophical advantage is that you can peek as often as you like. The third major advantage is that you can alter your test material in the middle of the test».