Originally posted by nightmarex
View Post
If from 5 million users you pick 10000 at random, that's *perfectly* representative, as it's a big enough sample to average out any individual inconsistencies. You got 1 for Linux and 3 for Windows, someone else got 3 for Linux and only 1 for Windows.
Also, *if* you used the Windows version more often than the Linux one, then the 1/3 ratio is actually representative and reflects your higher use of the Windows version compared to the Linux one. In other words, if the most popular games are Windows-only (like Skyrim), then it's natural that the Windows version shows even bigger numbers, since you can't use the Linux version to play them.
Comment