Saber-Terms: Catcher Defense
By Michael Jong
Perhaps I should have gone into this before embarking on the WAR series, but this topic is relevant to measuring the value of players. Catcher defense is interestingly one of the least known quantities in baseball despite undoubtedly being one of the most important. I don’t think anyone, saber-slanted or otherwise, doubts the value of having a good backstop behind the plate. Catchers affect a lot of the game, and it’s a shame that right now, sabermetrics just is not prepared to put a number on everything that a catcher has his hands in. But, of course, we can try.
The catcher has a few responsibilities which we are aware of:
– Control of stolen bases
– Blocking pitches at the plate
– Fielding errors
– Framing pitches
Each is important, and each is somewhat difficult to quantify or credit. But as I said, let’s give it a shot. (Note: I didn’t develop any of these ideas. Credit should go to Justin Inaz of Beyond the Box Score and other predecessors for much of these methods)
Stolen base data is the primary way that catcher defense has been measured because it’s some of the most available data out there. For years now, we’ve been judging catchers by their caught stealing percentages (CS%) because it’s really the only number we’ve had available to us. Applying these numbers to runs isn’t all that difficult.
Take a catcher’s CS% and subtract the league average rate. That gives you the number of extra CS your catcher gets per stolen base attempt. Multiply that by the stolen base attempts on that catcher and you get the amount of extra CS that catcher got over an average catcher. From there, all we need to do is apply a run total, the difference in runs between a CS and a stolen base. Here are how the Marlins’ catchers fared, with the run difference being 0.62 runs.
John Baker: 20% CS%, -5 runs
Ronny Paulino: 31% CS%, +1 run
Those are round to the nearest whole run to reflect the accuracy of defensive stats in general. The league average CS% here is 28%. We can also do this based on CS from catchers only, thus removing CS caused by pickoffs. Doing this gives a league average of 23%, and it yields these runs above or below average for Baker and Paulino.
John Baker: 18.5% CS%, -3 runs
Ronny Paulino: 27.5% CS%, +2 runs
Either way, you can see the general trend, Paulino was solid at catching runners stealing, while Baker was not. But, as we know, the Marlins’ pitching staff did no favors to their catchers in their slow deliveries. A big concern during the season was how the Marlins’ pitchers weren’t doing their job in holding runners at bay. And that gets to the problem of this methodology: this method gives full credit to catchers for this part of the running game when clearly pitchers and catchers share in the responsibility. Even stolen base defense is not without its problems.
With the data available at Baseball-Reference, blocking pitches is now also much easier to access. Matt Klaassen did a piece on Driveline Mechanics (before he left for bigger and better things at FanGraphs) on catcher evaluation, and I’ll use the methodology he uses there. Essentially, find the rate of wild pitches and passed balls per plate appearance caught allowed by your catcher and the league, subtract the league rate from your catcher’s rate, and multiply by innings caught. Then we apply a run value of a passed ball/wild pitch, which Klaassen calculates at -0.28 runs. Here are the Marlins’ catchers rates per PA caught and runs above or below average. The league average is 0.0097 WP+PB / PA.
John Baker: 0.0102 (WP+PB)/PA, -0.5 runs
Ronny Paulino: 0.0103 (WP+PB)/PA, -0.5 runs
Both Paulino and Baker were a bit worse than average at preventing pitches from getting by them. My mind recalls that Baker was not very good at this, but apparently he wasn’t all that much worse.
As Klaassen mentions in his article, errors are also included in B-R, and we can approximate error runs in the same fashion as we did blocking pitches. The error rate per PA was 0.0002 fielding errors/PA and 0.0011 throwing errors/PA. The run value for a throwing error was -0.48 runs, while the fielding error was -0.75 runs.
John Baker: 0 runs
Ronny Paulino: 0 runs
Neither player did much to hurt their status, but error runs on catchers are always going to very marginal.
Framing research is now possible, though it is in its infant stages. Thanks to Pitch f/x, we can find locations where 50% of pitchers are called for strikes or balls. With that data, we can check how catchers do in those regions and credit them for any extra strikes they get called. The difficulty is in finding the appropriate regions. Strike zones vary drastically depending on hitter height and even on individual umpires, so controlling for these factors may be difficult. To get an idea of the discussion right now, here’s BtB colleague Dan Turkenkopf’s attempt at using Pitch f/x data to get started on the debate, and here’s Lookout Landing’s Matthew Carruth’s attempt at framing that framing debate, along with Tango’s Book blog post discussing it.
This is one of the most controversial pieces of catcher defense that is often referenced. We know that catchers call pitches, but is there any way to determine how much better a catcher is at calling pitches than another one? Can we put a number on how well a game is called? With Pitch f/x, we can see how a game is called, but obviously we don’t know how much of the results of those calls are the catcher’s choice in sequencing and location and how much of it was the pitcher’s execution. The one thing we do know for sure is that we should not ever use catcher ERA. Not only are there significant problems with using ERA as an evaluator of pitchers because of defensive effects, but using it for catchers also doesn’t correct for a pitcher’s individual skill set. In short, it is useless, and you should never use it.
It’s a shame that catcher defense is so ambiguous at the moment, while its game impact is as large as it is. The tools of ignorance really do get underrated. With Pitch f/x now available, we are getting closer to coming up with methods to improve catcher defensive evaluation, but we are not quite there yet. Personally, I’m very excited to see what new things come out of the framing discussion, and how that will give us new insight on the good and bad of catcher defense. I’ll leave you with the total runs that I found from Baker and Paulino last season.
Baker: -3.5 runs
Paulino: +1.5 runs
Both players have their value changed a little, but not enough to significantly affect their excellent offensive value from their position. As Marlins fans, we should be happy to have a productive set of players like these two donning the mask and taking their beatings behind the plate.