Wasn’t that a Conan Doyle story?

The analogy to players is just silly. Players’ performances change a great deal over the course of their careers. Ballparks (that don’t change their dimensions) do not. I think even a 3 year park factors are not a very large sample size for what are not generally huge differences between parks. Also, yes of course weather is important for park factors and is probably even more important than park dimensions. But weather patterns at a particular locale do not change, at least in human time. If some park was unseasonably warm or cold relative to its history in, oh, 2006, why on earth should some player who plays in that park get dinged or benefited in terms of OPS+ for his performance in 2008, if 2006 was a weather-related aberration?

Where this comes from for me is that OPS+ is a great, easy stat to reference, but using 3-year park factors in the A’s case makes yty same-team ops+ comparisons a little silly. The Coliseum park factor used for calculating OPS+ changed dramatically in each of 2005-06-07. But the Coliseum didn’t change one lick! Swisher had a higher OPS+ in 2007 than in 2006 despite opsing 30 points higher in 2006. Most of that was the decline in average AL offense, but to whatever extent it was based on different park factors for 2006 and 2007, it is completely absurd. The effects of park dimensions and weather are very important but also take place on the margins, and I would not trust for a even a second a 3-year park factor done rigorously with no other evidence.

That said, bbref’s park factors are the probably the best we’re going to do, since new parks and changed dimesnsions are common, unless someone puts in a lot of effort every year.

(I am 95% sure that the dimensions of Turner Field haven’t been changed since it was built.):
This guy:
http://www.andrewclem.com/Baseball.php
has great current and historical park info, with graphics of changing park dimensions, etc.