Bias towards right frame

[Jon Seymour]

G'day,

This is a great tool - excellent for doing simple A/B tests.

One comment is that for heavy pages, I have observed that the results tend to be biased towards the right frame.  This is most easily seen by loading the same URL into both frames. 

I saw this is on Chrome (Version 27.0.1453.93) running on Ubuntu 12.04.

One way to counter act this might be to swap the frames on each repeat test so that neither frame gets an overall advantage from whatever bias is being introduced by the browser.

jon.

[Jon Seymour]

On Sunday, 16 June 2013 13:39:32 UTC+10, Jon Seymour wrote:

One way to counter act this might be to swap the frames on each repeat test so that neither frame gets an overall advantage from whatever bias is being introduced by the browser.

Ok, I see you have thought of this already with the serial option. 

Anyway, thanks for the tool.

jon.

[Ryan Witt]

Yup, good idea. I think I tried that at some point in development and opted for serial mode for more serious data gathering. 

Mostly, whichloadsfaster is intended as an evangelism tool, pointing people to other services and hopefully leaving people with the impression that performance matters. It would be better if it measured the time to render content before the fold, but that gets tricky and harder to do in the client side.

Thanks for the kind thoughts! Glad you like it.

--Ryan

--
You received this message because you are subscribed to the Google Groups "whichloadsfaster" group.
To unsubscribe from this group and stop receiving emails from it, send an email to whichloadsfast...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.