That does assume that the 1 in 20 which perform poorly are randomly distributed across the set of PDFs. It seems at least plausible that these would actually be

That does assume that the 1 in 20 which perform poorly are randomly distributed across the set of PDFs. It seems at least plausible that these would actually be grouped up by type somehow (e.g. it mentioned that the worst one they found was a huge vector map of the Lisbon subway system), and thus it would probably be the case that a user would either encounter them far less than that 1 in 20 or far more, depending on their own usage patterns.