One of the niftier aspects of the team defensive metric that I am now in favor of is that it requires only two pieces of data. I need the result of the play, which is easy to get and mostly objective. I also need the batted ball type, which is tougher to get and also murkier. The foggy lines of delineation in batted ball types I have dealt with as best I can, and will continue to tweak, but sheer access to them limits how far back I can look in baseball's past.
Luckily, Retrosheet has data on batted ball types stretching from present day back to 2003 and, thanks to Project Scoresheet, again from 1999 to 1988. With a few code modifications, I am able to apply the exact same algorithm and equations that I used on MLBAM data to create the fielding rankings to my Retrosheet data and look backwards.
That is what I am currently compiling, but a thought struck me as I got to 1999 and I decided to break that data in half around the All Star Break and poke through it. See, that is when the Mariners moved mid-season from the Kingdome to Safeco Field. Not many teams make a stadium move mid-season, and doing so allowed me to compare numbers from the first and second half more directly than I could, say 1998 to 2000 or between the 2009 and 2010 Twins.
The other nice part about the 1999 Mariners is that their starting position players did not change much over the course of the season, which helps makes these comparisons more useful. David Segui being traded away for Tom Davey and Steve Sinclair at the July deadline was about the only alteration on the team. Right field didn't have a dominate starter, but the infield of Dan Wilson, David Segui (afterward replaced by a hodgepodge), David Bell, Alex Rodriguez and Russ Davis was almost entirely intact from start to finish. That's important because one number in particular jumped out to me.
1999 Mariners' defensive RBBIP* on ground balls
*Reached base on ball in play. Defined as BABIP but including errors.
The departure of David Segui probably had some effect here, though I do not even know if it was positive or negative. Random fluctuation is also at play, but we are talking about decently large samples, about 1,000 ground balls in each half. What did change was the playing surface.
I doubt switching from Astroturf in the infield to grass was responsible for a 50-point RBBIP dropit had that large of an impact. Road games are included in the sample here and I would need to investigate other teams that made that switch for instance to get a feel for how much might be reasonably attribute to Astroturf. Nor is Astroturf being tougher to field ground balls on a remarkable hypothesis to utter. However, I had not seen it put into a quantifiable number before and though this is merely one solitary data point, I find it interesting for its possible implications.
[UPDATE]: I realized the categories of Kingdome and Safeco Field were potentially misleading. I was not isolating the .251 and .303 rates to ground balls fielded at those respective stadia, but rather during the period of play (pre and post All Star Break) when those were the Mariners' home park. That is, it included the Mariners' defensive performance on the road as well. I have changed the categories to hopefully remove any distinction.
I went back and calculated the numbers only at home. Obviously a smaller sample but at the Kingdome, the defense posted a .308 RBBIP on ground balls while at Safeco that rate improved dramatically to .245. Inversely, Mariner hitters at the Kingdome had a .272 RBBIP on ground balls in 1999 and a .253 RBBIP once moving to Safeco Field. Clearly, the playing surface is not solely responsible as the Mariners' defense improved in the second half in road games as well, from .299 to .256, but there does appear to be roughly a 20-point gap that might be on the turf.