TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Start here: Statistics for A/B testing

45 点作者 guico超过 9 年前

2 条评论

yummyfajitas超过 9 年前
I like this article a lot. But there is one thing that it gets a bit wrong.<p>The article is discussing the standard textbook Z-test. The article then talks a lot about Optimizely. However, Optimizely doesn&#x27;t actually use the Z-test - they have a sequential testing method instead, and the details are a bit different.<p>The article also suggests &quot;start by serving variant B to only 10% of the users to ensure there are no implementation problems&quot;. This is a good idea, but once you&#x27;ve ensured there are no integration problems you need to throw away the data and restart. Since conversion rates change during the week (i.e., sat != tues), keeping the data during the ramp-up period is a great way to get wrong results due to Simpson&#x27;s Paradox.
评论 #11041058 未加载
ep103超过 9 年前
Would love a non-optimizely, guide to building proper A&#x2F;B tests, if anyone knows one
评论 #11041757 未加载