A few months ago, one of my favorite Pitch F/X writers Jon Roegele posted what basically amounted to Pitch F/X park factors. I have used them in some of my articles before, but as he notes in the article, in his actual table when it came to velocity, he didn't control for pitch type (he notes the problem with this in the article, and does say that looking at just fastballs does change the data and gives some examples), so I wanted to do my own version, just looking at velocity in 2013. Unlike Roegele, I did not control for handedness.
The first thing I did was look at all 4-seam fastballs (FF in the classifications) for each team based on the home and road. This has some problems of its own, as we are dealing with MLBAM pitch classifications, which certainly are not perfect. However, I didn't want to include all pitches or all kinds of fastballs because we aren't controlling for pitch type, which could skew the velocity if a few pitchers changed their tendencies from home to road, of which there seem to be some cases. Of course, if there are tag biases in parks, it could affect the data in a way we can't control. There is also the problem of roster fluidity. Teams use different pitchers throughout the season, releasing and signing guys, using spot starters, having to put pitchers on the DL, etc. So some pitchers may pitch more on the road than home (this especially applies to spot starters). Depending on their velocities, this could mess with the data (also, sometimes it just works out where some pitchers pitch more on the road than home or vise versa). To control for that, I am putting two sets of data, the team home/road splits, along with the home/road splits of the pitcher on each team with the most fastballs thrown. I think I did it a little differently than Roegele, looking at each team individually, splitting the teams by home/road (and not separating by road parks). Baseball Savant made this really easy to do, so this is where I pulled the data from.
Of course, this data will instantly out of date, so don't use the raw MPH data. This is only to get a better look at which parks have biases in velocity data. The differences are far more important than the actual velocities. I sorted the spreadsheet by the overall team difference.
Weirdly, it seems that the average velocity of all pitchers is slightly higher on the road. Obviously the bigger differences came from the individual pitchers, which would be expected due to sample size. The Blue Jays have the biggest difference when it comes to team, with Justin Masterson having the biggest home/road splits out of the 30 pitchers. The Braves results are really weird, as Mike Minor has an average velocity of nearly half a MPH on the road, but the park actually is the 2nd "hottest". The Royals park, usually thought of and tested as "hot" has played neutral as a team, but Ervin Santana has seen a big difference from home to road. Overall, the results for all parks aren't that far off on average, and I don't think it really changes much when it comes to analysis of individual pitchers.