Add new comment

Honestly, the Pittsburgh stats make perfect sense to me.  In 2006, he had 316 POs in 1373 innings, (2.1 RF/9), and a 316 PO versus 307 expected out comp.  UZR had his range at plus 3.3.  I can buy that.
By 2007, the righting was on the wall that Pitt wasn't going to be getting better while Bay was with the team.  The 2007 season, Bay likely played hurt for much of the season, (it was BY FAR), his worst offensive season, also.
In the end, he made 265 POs in 1237 innings, (2.0 RF/9).  Per FG, he should've made 280 outs, but only managed 265.  So, his -8.9 range factor isn't outrageous.  But, he WAS slowed by injury for some of the year -- and he likely also went thru the motions for much of the year.  His career high 8 errors supports that 2007 was a bad year in EVERY aspect of his game.
In 2008, he managed only 254 POs in 1344 combined innings.  That *IS* a very poor showing, (and a personal worst 1.8 RF/9).  By 2008, he just wanted out of Pirate purgatory, and simply concentrated on hitting -- and likely became Ichiro-like in LF, not wanting to risk an injury costing him a chance to move somewhere.  And he did move -- into the park with perhaps the hardest LF in the world to play.  If there's any LF on the planet I would expect a learning curve, it would be Fenway.  He also managed only 8 assists in 2008, (his career worst for 150+ game season).  So, the combo of his lowest outs - his fewest assists since 2005 ... and yes, his 2008 UZR can be defended.
My point is that the 2009 UZR *CANNOT POSSIBLY BE DEFENDED*.  He made 310 POs in 1279 innings -- setting a personal best in RF/9.  He had a career high 15 assists.  He had a career best ZERO errors.  Okay, I can appreciate that if you have a weak arm, teams will run on you, so often the best arms have low assist totals.  But, Bay led the majors with 15 assists from LF, and cost his team 0.7 runs ... while DeJesus threw out 13, and saved his team 6.7. 
My problem with UZR, is that it say Bay *should have* made 25 more outs than any other LF in all of baseball.  Carlos Lee, for the Astros, played only 7 innings less than Bay.  According to this system, in Houston, the "expected" outs was only 217.  In Boston, 327.  I am supposed to believe that somehow, in Boston, there were 120 more "gettable" balls in LF than in Houston.
In looking just at the Boston 2008 stats: Manny had 99 POs in 537.2 innings, (1.8 RF/9), while Bay was just as bad (76 POs in 432.1 innings (1.7 RF/9).  Yet, Manny was pegged at only -3.0 runs, while Bay was pegged at -8.8 range runs cost.  Their RF values are really, really close - (and both significantly behind Ellsbury).  But, Manny, who had 116 more innings cost the team half as many runs via range?!?.  Manny's 6 assists in his 66 games netted him a 4.5 arm, while Bay's 5 assists in 49 games got him 0.4.
It just doesn't track.  I don't believe there is some cosmic field surrounding Jason Bay that causes an exponentially larger number of balls to be hit in his direction that would be catchable by an average LF. 
What I believe is that Bay "gave up" in Pittsburgh, and his poor range ratings are supported by his poor PO numbers.  I believe in 2009, in Boston, his defensive results improved dramatically, (playing for a contender often has that effect), and his raw countables jumped drasitcally.  But, UZR and RZR don't show that.  Why?  Because *ZR* is the problem.  It's the foundation block for most of these stats - and if ZR is wrong, every variation of it is wrong.
Bay jumped from 254 outs in 1345 innings to 310 outs in 1279 innings, and UZR said his range improved from -14.6 to -14.4.  That's 56 extra outs in 66 fewer innings.  Think about that for a minute.  In 7 FEWER games, he netted more than two complete perfect games worth of outs.  UZR is suggesting that to be AVERAGE for Boston in 2009, a LF would've needed to have roughly a 2.5 RF.  In the past 5 years, Carl Crawford is the only LF in all of baseball to post a 2.5 RF. 
You go back to 2005, when Manny only managed 243 POs in LF (1225 innings - 147 games), I can at least entertain the notion that Crawford, (in similar innings), who had 341 POs, was really 35 runs better than Manny in range.  341 POs to 243 POs ... in similar innings.  Yes, I can accept that there could be a major difference in run prevention.
But Bay, in 2009, was second only to Crawford in putouts, had a basically identical RF/9, and they played the same innings, (3 apart).  I CANNOT accept that the 17 extra POs Crawford managed in 2009 explain the 32 run differential in range rating from UZR.  It makes no logical sense.