Effective A/B Testing

Posted by luu on Tuesday May 19, 2015 @05:58AM
from the elem.com dept.

http://elem.com/~btilly/effective-ab-testing/

( 5 comments | 11 points | on HN )

Re:Effective A/B Testing by bgun on Tuesday May 19, 2015 @02:30PM (#9572140)

There may be useful information here but the presentation format is incomprehensible. And do the images on the right bear any relationship to the content?

[ Parent | on HN ]

Re:Effective A/B Testing by caminante on Tuesday May 19, 2015 @01:43PM (#9571810)

Is there a way to advance by complete slides and not have to click for every intra-slide transition?

It took 47 clicks to get to slide 11 and the deck has 115 slides.

[ Parent | on HN ]

Re:Effective A/B Testing by byamit on Tuesday May 19, 2015 @03:56PM (#9572695)

http://meyerweb.com/eric/tools/s5/features.html#controlchart

A couple useful shortcuts there depending on what you're trying to do.

[ Parent | on HN ]

Re:Effective A/B Testing by caminante on Wednesday May 20, 2015 @03:04PM (#9578429)

I saw that. Typing '1'+<RIGHT ARROW> advances slides, but doesn't load the full slide.

Bad presentation design.

[ Parent | on HN ]

Re:Effective A/B Testing by Matumio on Wednesday May 20, 2015 @04:06PM (#9578917)

There seems to be a lot of experience behind this presentation. But I wonder, how does all this G-testing compare to Thompson sampling? (Which was not mentioned at all in the presentation.) For example, is there any drawback with adapting the A/B probabilities on-the-fly, instead of having a hard decision point?

[ Parent | on HN ]