Whether it be the weather in Florida or the way the infield was groomed, I concur that the condition of the home ballpark’s infield is the the obvious source of statistical oddities like these. Your research on home and away ground ball rates seals it for me, even if the data is merely suggestive, not definitive. Kudos.

Is it possible to look up home-road fielding splits without downloading game files from Retrosheet and doing the analysis one’s self? I’ve often been interested in looking up such information, and I’ve been unable to find much.

In any case, great post to a good article (thanks, Eric!)

Thanks.

Uggla and Jacobs – The first striking thing about Uggla’s stats is the number of balls in his zone fielded by the first baseman.
—–1B_Plays—SS_Plays
2006—7——–7
2007–28——–1
2008–15——–8

Even though Jacobs was worse overall by UZR in 2007 he went from being 9 plays below average on balls in the 2B’s zone in 2006 to being 2 plays above average in 2007. Most of the additional 21 plays fielded in front of Uggla were probably balls that he could have easily fielded for outs. But instead 8 of the balls fielded by the 1st basemen were infield hits. BZM treates all balls fielded by the first baseam in front of the 2nd baseman as non plays for the 2B, but UZR docks the 2B value if the ball is not fielded by the 1st baseman for an out. So this one area alone would contribute several negative runs to Uggla’s 2007 UZR.

Why this big change for Jacobs in 2007? First of all Jacobs only played about 2/3 of the innings each year and had a different backup each year for the other third. Helms was the primary backup in 2006, Boone in 2007 and Cantu in 2008. Cantu was the only one with above average range. Helms a bit below average and Boone even worse than Jacobs.

The second set of numbers in the table is the help that Uggla got from the SS on balls up the middle. 2006 and 2008 are a run or 2 above average help and 2007 is 5 plays worse than average. This is probably an positioning adjustment by Ramirez to try and compensate for Cabrera’s poor fielding. The bottom line is that Uggla is about an average fielding 2B whose numbers were inflated by those around him and chance in 2006 but were worse than normal because of those factors in 2007.

Ramirez and Cabrera/Cantu – BZM actually has Cabrera’s LW/150 drop from 4 below average in 2006 to 20 below average in 2007, much greater than UZR. This is almost all due to a drastic decrease in range, both in and out of his zone. In zone went from 8 plays below average in 2006 to 31 plays below average in 2007. Out of zone went from 8 plays above average in 2006 to 8 plays below average in 2007. Although BZM and UZR both have Cantu below average overall in 2008, BZM says it wasn’t because of his range, which was 2 above average in zone and 3 above average out of zone. Cantu’s above average errors and below average in starting DPs put him below average overall. Ramirez is usually a bit above average in range but below average in errors but in 2007 his in zone range was terrible, 24 outs worse than expected. Cabrera did make 9 fewer plays in front of him but that and chance do not seem to be enough to explain this abberation. Those of you who follow the Marlins closely, was Ramirez playing with an injury in 2007? The bottom line is the Ramirez is a slightly worse than average SS overall that had an exceptionally bad year in 2007.

It could still be the pitchers. The Marlins did allow a higher percentage of ground balls to go the outfield overall in 2007, which doesn’t really say much since we already know the infield defense was worse.

Do you have data on defensive ratings to the players’ left and to the right? Perhaps the team’s defensive positioning up the middle was just bad in 2007, possibly as an attempt to compensate for bad corner defense. It may have been a misguided positioning effort rookie manager Fredi Gonzalez tried and then scrapped in 2008 when it didn’t work, but that’s pure speculation with no real support. Or maybe it’s just a complete fluke. It is very interesting how the numbers coincide, though, and I’d love to see more research into what happened.