zader

Joined 17 years ago

Karma
143
Submissions
26
Member since
Aug 3
CEO of Atlas Web Development (www.AtlasWebDev.com) and founder of the Atlasphere (www.TheAtlasphere.com).

joshua at atlaswebdev dot com

View on Hacker News

Recent submissions

Comment · 14 years ago
Good catch. I figured HN would reject it if it had already been posted. Was surprised when it went through, actually.
Comment · 14 years ago
Estonia down.
Comment · 14 years ago
What are the "big issues" that would be caused by forcing SSL? It sounds like visitors from Iran (and other countries that block SSL sites) might be unable to access HN. What else?

(To clarify, I'm interested in the general topic of whether it's a good idea to have all-SSL websites.)

Comment · 14 years ago
Comment · 14 years ago
Comment · 14 years ago
Comment · 14 years ago
Helpfully, the final paragraph of the article seems to contain its takeaway message:

"So, comparing A/B testing and multi-armed bandit algorithms head to head is wrong because they are clearly meant for different purposes. A/B testing is meant for strict experiments where focus is on statistical significance, whereas multi-armed bandit algorithms are meant for continuous optimization where focus is on maintaining higher average conversion rate."

Like any business owner, at the end of the day I care more about conversion rates than about statistical significance.

I'm not a scientist trying to advance the sum of humanity's knowledge. I'm a business owner trying to find the shortest path between what my customers need and what I can profitably offer them.

In a way, statistical significance strikes me as a bit of a fool's errand, because significant results in one context may not be generalizable to another, which means even if we know for a near certainty what worked best, it's hard to apply that knowledge reliably in the future.

Of course, with MAB we could still wait for statistical significance if we want it, before turning off variations that are performing worse. And we can certainly still try to draw conceptually useful conclusions by designing our test variations in ways that facilitate easy comparison.

But with MAB and Myna it sounds like we can pretty well count on higher conversion rates at the end of the day, as well, and that counts for a lot in a business context.

I'm grateful to the VWO team for writing up their analysis and findings, and being so frank about the relative advantages of A/B and MAB. Their summary above tells me what I need to know.

Comment · 14 years ago

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton