mkrizan86:I would only add 1 thing to Robert's screening process. One needs to take into account the number of red tee rounds, where putting stats and distance from the pin tend to get a huge boost.
Of course, you're correct. But I decided not to open up that particular can of worms.:-)
All of this is basic probability theory action. I guarantee, with that large of a sample size, that the number of putts/hole for each round, when organized, will form a normal distribution. Additionally, the range of possible values is actually quite narrow, limiting the possible effect of outliers.
What all of that means is that if one has enough current data, even with externalities considered, the mean is a meaningful and accurate number, representing what the player is actually doing.