Catcher Defense in Phoenix – A Study, Redux

[RonCo]

After talking with Tyler this weekend, and reading his study again, I got to wondering how catcher defense would manifest itself in the data I’m capturing with live BBA data. This is interesting to me because, while studies like the kind Tyler did do a great job identifying if a skill or trait actually exists, the magnitude of them in controlled environments is hard to determine. By that, I mean that the study Tyler ran was with all ratings neutralized except for catcher defense, and then catcher defense carefully crafted to force result to tell you something.

The live BBA data will be messier because the competitive environment isn’t controlled like that.

Anyway, I got to thinking about how I might use the BBA data to look at catchers. Here’s what I came up with:

1) Find a team that uses two catchers for a reasonable amount of time
2) And whose catchers have very different defensive ratings, ideally one very good and one not so good.
3) Compare pitching staff results, both as a whole and in individuals, when each catcher is being used

This, I figured, should give me at least a one-shot look at how big of an advantage great catcher defense is in the wild. After picking around a bit, I focused on Phoenix, who uses a pair of catchers:

• Cisco Arreola: 36 yo, 3/6 Ability/Arm
• Vicente Diaz: 25 yo, 6/8 Ability/Arm

This isn’t a mega-gap, but it’s a bit of one. At least Arreola is clearly aged out now, and Diaz can arguably be seen as a league average kind of defensive player.

I then had to do a little dipsy-do with the spreadsheet data, identifying which games and innings the two actually caught, then transposing those into the pitch/PA data. Once that was done, the rest was easy. I should note, I suppose, that the data I'll present here is captured through the end of August.


OVERALL/COMPOSITE REVIEW

First, I looked simply at strikes and balls, regardless of the pitcher. This is probably mostly a silly academic study. My guess is that OOTP assigns these after the fact, but (1) I don’t know this, and (2) I was having fun trying to see if a better defender got more called strikes—which is what framing is, right?

The bottom line is: yes, Diaz got more strike calls and had fewer balls called on the whole.

Catcher	Pitches	StCa	StSw	Balls
Cisco Arreola	14498	15.8%	9.4%	37.9%
Vicente Díaz	4079	17.0%	10.5%	37.3%

The half percent fewer balls may or may not be significant at this sample size. I haven’t checked. But it is in the general direction one would assume, and given that Diaz’s ratings aren’t elite, a smallish gap seems reasonable. (Note, too, I haven’t normalized anything for the sample of pitcher’s they’ve caught. Each catcher has caught each pitcher on the staff, but the ratios could be different, and that could nudge things one way or the other, too). The full percent or more of extra strikes is interesting, though.

Lets go the next step and look at results.

Are these distributions resulting in more strikeouts and fewer walks, and if so, how many?

The answer is: yes, a few fewer walks and considerably more strikeouts.

Catcher	PA	BB/PA	K/AB	KL/AB	KS/AB	HR/AB
Cisco Arreola	3895	7.8%	20.7%	5.0%	15.8%	3.3%
Vicente Díaz	1114	7.5%	24.4%	6.6%	17.8%	3.0%

Among the things I like here are that small gap in walks harkens back to the small difference in balls, and that the larger difference in strikes is reflected in a larger number of extra strikeouts for pitchers Diaz catches.

Again, though, this is composite data and could be influenced by several selection issues.

Note, also, that I’ve included HR/AB rates, though I have no idea why a catcher would influence those. Fun, remember?


A LOOK AT INDIVIDUAL PITCHERS

So, yeah, fun stuff. But there are so many degrees of freedom here, not the least is that a catcher’s performance in those numbers could probably be swayed heavily by which pitchers he caught most. One guy gets a fireballing strikeout guy, and the other guy catches the junk baller, and see these numbers go haywire.

So, instead, here’s a look at a few of Phoenix’s pitchers, focusing on the guys with bigger sample sizes. This will still be noisy for sample size purposes, of course, and will still be influenced on the fact that I haven’t put the results in context of what hitters comprised the sample. Guys who face more sluggers, for example, could have different numbers here than guys who faced a bunch of high-contact slap hitters. So when you look at the numbers, you have to decide yourself how far to trust them—including none at all.

Let’s look, for example at Charlie Iron-Knife. Iron-Knife. He’s a knuckleballing lefty with limited control who leads the team with over 700 batters faced.

Catcher	Pitcher	PA	BB/PA	K/AB	KL/AB	KS/AB	HR/AB
Cisco Arreola	Charlie Iron-Knife	596	9.9%	21.1%	4.6%	16.5%	2.1%
Vicente Díaz	Charlie Iron-Knife	137	11.7%	30.5%	8.5%	22.0%	1.7%

In this case, Iron-knife allows more walks with Diaz than he does with Arreola, but also considerably more strikeouts. One has to ponder sample sizes, and one has to ponder the influence of Iron-knife’s poor control, regardless.

Here’s 24-year-old power pitcher Jose Aguilar with 584 PA:

Catcher	Pitcher	PA	BB/PA	K/AB	KL/AB	KS/AB	HR/AB
Cisco Arreola	César Aguilar	427	7.0%	19.1%	3.6%	15.5%	2.8%
Vicente Díaz	César Aguilar	157	7.6%	28.0%	9.8%	18.2%	2.1%

Again, Aguilar walks guys a touch more often when Diaz is back there, but is striking guys out at a rate nearly 10% higher. Given randomness, Aguilar’s results seem to confirm the feel of the overall rates, too.

With Maxime Manceau, we come to a different profile though. In 700 AP:

Catcher	Pitcher	PA	BB/PA	K/AB	KL/AB	KS/AB	HR/AB
Cisco Arreola	Maxime Manceau	507	7.1%	15.3%	4.6%	10.7%	2.6%
Vicente Díaz	Maxime Manceau	193	3.6%	14.9%	5.5%	9.4%	3.9%

Manceau is an extreme junkballer who relies on command, control, and the well-timed grounder. He just really does not strike a lot of guys out. So it’s probably not too surprising that his performance under the two catchers is about neutral when it comes to strike outs. With Diaz, though, he’s walking guys half as often as with Arreola.

Here’s Elwood Blues, whose data seems to support Diaz across the board. Note, however, that sample size. Arreola seems to be Blue’s primary receiver.

Catcher	Pitcher	PA	BB/PA	K/AB	KL/AB	KS/AB	HR/AB
Cisco Arreola	Elwood Blues	512	9.8%	21.2%	6.3%	15.0%	4.2%
Vicente Díaz	Elwood Blues	41	7.3%	27.8%	5.6%	22.2%	2.8%

Walks down, K’s up.

Hector Amaral, too, appears to fare a touch better with Diaz behind the plate, though the impact looks subtle based on this data.

Catcher	Pitcher	PA	BB/PA	K/AB	KL/AB	KS/AB	HR/AB
Cisco Arreola	Héctor Amaral	497	5.6%	22.1%	4.8%	17.2%	3.7%
Vicente Díaz	Héctor Amaral	190	5.3%	25.3%	5.1%	20.2%	0.0%

At this point, we start getting to guys with sample sizes down in the ranges that are so low as to be make-work to post them here.


SO WHAT’S THE ANSWER?

I hear you. I started this to get an idea of how big the impact of catching really is in the wilds of the BBA. And in this case, I’ve got a data set that sets a poor defender (Arreola) vs. a competent one (Diaz). What, you ask, do I think the answer is? How much of an advantage do Phoenix pitchers get when they throw to Vicente Díaz rather than Cisco Arreola?

The answer is: I don’t know.

I’d baseline my thoughts on the idea that they’ll walk a half-percent fewer walks is probably good—maybe low. Depends on the pitcher. I’d also say that it wouldn’t surprise me to see if a bigger study said pitchers throwing to Diaz might strike out an extra guy every 3 games or so. I note, also, that in this very small sample it almost appears that Diaz helps pitcher give up fewer homers—though that might be due to the pitcher just striking out more guys. Dunno.

And I’ll note that I didn’t get a sample size from pairing an elite level catcher and a mid-range or worse partner. It’s certainly possible that the influence is small on the low end, and ramps up when an elite level catcher is in there.

Regardless, as I noted to Tyler in our discussion, the most important thing about these studies is the degree to which they confirm or deny that you can use “normal GM logic” (whatever that is) when assessing values of players, and what I’m seeing here does seem to line up with Tyler’s study that shows catcher defense in OOTP is something that influences the game in more ways than just controlling the running game.

Which is pretty cool, right?

[crobillard]

This is pretty interesting. Is the thought here that catchers with higher ability ratings are better at framing pitches or perhaps just making the pitcher better at what they are good at? I'm not really sure what to make of the data. I always assumed catcher defense mattered, but I just assumed it meant less wild pitches, more runners thrown out. I have Fabre and Katzenbogen, but I haven't played Katzenbogen much this year. Last year he appeared in 56 games to Fabre's 127, but I think that's still too big of a difference.

It's a good question though because if a catcher can in fact make a pitcher better at what they're good at than that could be something very useful. I wonder how many wins I had last season with Katzenbogen starting vs. Fabre. Interesting stuff. I wish I was smart enough to know how to pull and manipulate this information.

[RonCo]

As Tyler's work seems to show, and this seems to corrobrate (and other things we've done around her support) Catcher framing is almost certainly in the game. The question is how big is the influence, which will probably always be an unknown.

It may be interesting to note that in areas where pitchers have holes (low control/low stuff) this look at Phoenix suggests catcher difference is of less impact...which maybe makes sense.

[jiminyhopkins]

Wait, what?

[udlb58]

As Tyler's work seems to show, and this seems to corrobrate (and other things we've done around her support) Catcher framing is almost certainly in the game. The question is how big is the influence, which will probably always be an unknown.

It may be interesting to note that in areas where pitchers have holes (low control/low stuff) this look at Phoenix suggests catcher difference is of less impact...which maybe makes sense.

Like you said, catcher defense in my highly controlled world does not exactly equate to catcher defense in the "real" world of the BBA. But in my research, it was pretty significant. I didn't do any win calculations, but the K:BB rates were significantly different for a bad defender vs a great one.