Good. By the way, the time.sleep wouldn't even have worked -- the big deal with timing things on the web is that unless you run some kind of autobanning script,

Good. By the way, the time.sleep wouldn't even have worked -- the big deal with timing things on the web is that unless you run some kind of autobanning script, they get unlimited attempts. I'm pretty sure that if I get a hundred-thousand-similar-attempt dataset I can calculate the mean time to such accuracy that the random sleep time doesn't hamper me at all. And I can collect that data over such a large time you might not even notice.