What You Need to Know:... (03/31)
4 comments have been left for this article. BP Comment Quick Links
| |
|
|
March 31, 2014 Framing the FutureCatcher Framing Projections for 2014One of the benefits of our recently released catching defense metrics is they’re essentially ready-to-project, thanks to the regression feature of the model (the "R" in RPM). RPM also gives us two ways to assign value to framing, one using context (the ball-strike count) and one using a flat value (recently adjusted* to ~.155 runs).
To date, we haven’t found an aging curve in the PITCHf/x-based framing metric we developed. This makes projecting framing performance a little easier, and for extreme simplicity I have created a set of projections for 2014 based on a 3-2-1 weighting of framing opportunities for the past three seasons. First, though, we wanted to generate a "retrojection" to test RPM’s success in predicting 2013’s framing performances.
Since RPM seasons are regressed to careers, and pitcher and umpire adjustments are derived from those actors’ “career” numbers, a retrojection requires recreating the 2008-2012 numbers as if 2013 had never occurred. Since the probabilistic portion of the model (the "P" in RPM) is specific to each season, there's no need to recreate the zone maps (see first footnote below).
The WOWY (With or Without You) corrections for pitchers and the umpire corrections were all recreated without including 2013, and new regressed “career” lines were created with 2008-2012 totals for each catcher. Those 2008-2012 seasons were then regressed to these new, abbreviated career lines. Finally, the 3-2-1 weighting was applied for 2010-2012 to arrive at retrojections for 2013.
The verdict? The retrojections did very well. Both context-dependent and generic run projections (as rate stats) correlated with actual 2013 performance at better than 0.81. That’s a lot of variance (>65 percent) captured by a simple model.
So with a fair amount of confidence in the crystal ball and the playing time assigned to each catcher in our own Depth Charts, we can project player and team framing contributions for 2014.
Yasmani Grandal is currently projected to top framing magician Jonathan Lucroy in 2014, based on playing time projections from March 30 (which may be optimistic for Mr. Grandal, as the Padres recently added a third catcher, Adam Moore). It’s a close race, perhaps too close to call.
Look down just a few spots, and you'll see some other young backstops, along with the usual suspects (Brian McCann, Russell Martin, Ryan Hanigan, the Molinas). Yan Gomes just got a contract extension that raised some eyebrows; maybe his framing projection will help lower a few of them.
Grouping by team, we can see Tampa’s dynamic duo leading the way, with the Rockies' terrible trio bringing up the rear.
*Since first publishing RPM, we've reset the run expectancies used in the final calculations to vary by season, to reflect the drop in scoring we’ve seen during the PITCHf/x era. During that process, we realized that we were about .01 runs low on the generic framing factor. It’s a zero-sum** game, so you won’t notice a difference of any substance. **Next, we'll have to ensure that the framing totals zero out by season. The run totals do zero out overall, but in 2008 everyone seems to be getting penalized an extra run, with things moving in the opposite direction in 2013. This is reflected in our 2014 projections, too.
Harry Pavlidis is an author of Baseball Prospectus. Follow @harrypav
|
Interleague Report: Week One
The Atlanta team based rating scared me, but I think it's largely driven by doumit. God willing he doesn't spend much time back there.
Very interesting.
I'm a huge advocate for taking away Doumit's catching gear.