I am a bit curious where you came up with the formula for the standard deviation for a players hourly win rate. It seems to me that you have a type and have 1/n where you should have 1/T. The consequnece of the error is the result you will get could be quite a bit higher than it really is.
I went through the trouble of deriving the formula for the standard deviation for a players hourly win rate and arrived at
This differs from the formula in the article by replacing 1/(total time) on the outside with 1/n. There is no way that I can see to rectify these results. I can only conclude that either a) the forumla was mis-typed or the source of the original formula was in error.