[milkman]

Am I the only one who appreciates the work he did to bring us these findings? I've tried Cloudflare before and "underwhelmed" would be as good a description as any. I don't know if it helped at all.

If anything, this post confirms it.

[andypants]

The work? He didn't bring us any findings. He brought us ambiguous 'load times' with barely any information except that the numbers under the cloudflare section are sometimes slightly larger than the numbers under the non-cloudflare section.

Actually, I just took another look at the post. The lack of detail is because his 'tools' are just free web services like pingdom. There are no more details about his method on the blog because nobody knows how these businesses make their measurements, as they are closed source.

Exactly what is included in the load times? How many values are aggregated in the load times? Are the times averaged or aggregated in some other way? Were the websites consistent in terms of cache in each of the measurements? Etc.

Scientific method, please! Or at least, a bit more effort.