Okay so yes the number of proxies is 443; it would have been nice to analyze pretty much any other number.
It was probably to catch the reader's eye - 500 proxies wouldn't have jumped out quite as much (to me, at least). Also, had they omitted the number altogether, the headline would be making a much more radical (and possibly incorrect) claim.
to be honest it was a coincidence. I was pasting IPs from random proxy lists on the web and this was the number I found. I just wanted to have more than 400 proxies to test
21% supported HTTPS. 'Only' somewhere between 16.6% and 25.1%, depending on the amount of overlap between the 'Modified JS' and 'Modified HTML' categories, did the shady thing of modifying returned content.
I thought the point of the article was that even if they didn't do anything to your content, they could still be collecting quite a bit of your information if everything you did was over HTTP and not HTTPS.
Another experiment around a spied free proxies from Chema Alonzo (DefCon).