A/B Materials for In-Depth Study
A comprehensive list of articles and links to deep dive in online experiments!

Recently, I got involved in the adventure of A/B testing. I decided to go beyond simple math and online calculators and dig deeper into this field. In this journey, I found many useful articles, papers, researches and so on. If you, like me, want to become professional in A/B testing, then check the following links in this post. You’ll find many exciting things.
<div class="
image-block-outer-wrapper
layout-caption-below
design-layout-inline
combination-animation-none
individual-animation-none
individual-text-animation-none
" data-test="image-block-inline-outer-wrapper">
<figure class="
sqs-block-image-figure
intrinsic
" style="max-width:2500px;">
<div class="image-block-wrapper" data-animation-role="image">
<div class="sqs-image-shape-container-element
has-aspect-ratio
" style="
position: relative;
padding-bottom:57.84000015258789%;
overflow: hidden;-webkit-mask-image: -webkit-radial-gradient(white, black);
">
<img data-stretch="false" src="https://images.squarespace-cdn.com/content/v1/5e6d2f92970b8e6e6a8d1910/1584220062511-5NLE2GB6VSOS6SLZ9BUT/image-asset.jpeg" data-image="https://images.squarespace-cdn.com/content/v1/5e6d2f92970b8e6e6a8d1910/1584220062511-5NLE2GB6VSOS6SLZ9BUT/image-asset.jpeg" data-image-dimensions="2500x1446" data-image-focal-point="0.5,0.5" alt data-load="false" elementtiming="system-image-block" width="2500" height="1446" sizes="(max-width: 640px) 100vw, (max-width: 767px) 100vw, 100vw" style="display:block;object-fit: cover; width: 100%; height: 100%; object-position: 50% 50%" onload="this.classList.add("loaded")" srcset="https://images.squarespace-cdn.com/content/v1/5e6d2f92970b8e6e6a8d1910/1584220062511-5NLE2GB6VSOS6SLZ9BUT/image-asset.jpeg?format=100w 100w, https://images.squarespace-cdn.com/content/v1/5e6d2f92970b8e6e6a8d1910/1584220062511-5NLE2GB6VSOS6SLZ9BUT/image-asset.jpeg?format=300w 300w, https://images.squarespace-cdn.com/content/v1/5e6d2f92970b8e6e6a8d1910/1584220062511-5NLE2GB6VSOS6SLZ9BUT/image-asset.jpeg?format=500w 500w, https://images.squarespace-cdn.com/content/v1/5e6d2f92970b8e6e6a8d1910/1584220062511-5NLE2GB6VSOS6SLZ9BUT/image-asset.jpeg?format=750w 750w, https://images.squarespace-cdn.com/content/v1/5e6d2f92970b8e6e6a8d1910/1584220062511-5NLE2GB6VSOS6SLZ9BUT/image-asset.jpeg?format=1000w 1000w, https://images.squarespace-cdn.com/content/v1/5e6d2f92970b8e6e6a8d1910/1584220062511-5NLE2GB6VSOS6SLZ9BUT/image-asset.jpeg?format=1500w 1500w, https://images.squarespace-cdn.com/content/v1/5e6d2f92970b8e6e6a8d1910/1584220062511-5NLE2GB6VSOS6SLZ9BUT/image-asset.jpeg?format=2500w 2500w" loading="lazy" decoding="async" data-loader="sqs">
</div>
</div>
</figure>
</div>
Overview of the A/B testing field
These links are great to refresh your knowledge, remember the basic formulas, and to learn common mistakes and solutions for them.
Advanced Topics
Sequential Analysis
Peeking Problem
How Not To Run an A/B Test — Evan Miller with wrong conclusions and great response to it A/B Testing Rigorously (without losing your job) (and extension A/B Testing With Limited Data)
Peeking at A/B tests: continuous monitoring without pain | the morning paper
Bayesian
Bayesian vs Frequentist A/B Testing (and Does it Even Matter?)
Discussion on Reddit — Frequentist or Bayesian AB Testing Methodology? : statistics
Bayesian vs. Peeking Problem
Is Bayesian A/B Testing Immune to Peeking? Not Exactly — Variance Explained
Bayesian AB Testing is Not Immune to Optional Stopping Issues | Analytics-Toolkit.com
A/A
More
In addition to authors that wrote previous papers and articles, I recommend you to check out these resources:
Ronny Kohavi is a Microsoft Technical Fellow and Vice President of Analysis & Experimentation. You must explore his project ExP Platform. Also look for his recommendations on what to read.
Eytan Bakshy’s blog. Eytan is a senior scientist on the Facebook Core Data Science Team, who lead the Adaptive Experimentation group.
Evan Miller is a developer of statistical software. He has a series of A/B testing articles on his website.
Papers from the SIGKDD conferences. It’s a community for data mining, data science, and analytics.
The story “How Optimizely (Almost) Got Me Fired” and a paper about The New Stats Engine, where they fix the problems.
Companies Blogs:
Netflix — Quasi Experimentation at Netflix — Netflix TechBlog — Medium
Airbnb — Scaling Airbnb’s Experimentation Platform — Airbnb Engineering & Data Science — Medium
Booking — How Booking.com increases the power of online experiments with CUPED
Facebook — Efficient tuning of online systems using Bayesian optimization — Facebook Research
Yandex – Online Evaluation for Effective Web Service Development
Stitch Fix Technology - Suffering from a Non-inferiority Complex?