Batter's Box Interactive Magazine Batter's Box Interactive Magazine Batter's Box Interactive Magazine
Batter's Box reader, poster and econ-whiz Mike Moffatt has prepared a great article on the Jays' 2002 catchers, measuring their impact on the pitching staff. We are happy to bring this to you at Da Box.

The thrust of Mike's piece is that Ken Huckaby's gaudy catcher's ERA is a result of who he was working with, as opposed to what he was doing. Mike's method is something that would make an interesting study is broadened, and might tell us something about whether catchers really have an impact on the pitchers they hook up with.

If you have an article or other piece, feel free to run it by me (tybalt4@yahoo.com); we're delighted to employ pinch-hitters here (this is an NL-style website, not an AL-style website).

Over to you, Mike...



Catcher's ERA and the 2002 Blue Jays
Mike Moffatt

Over the last few months the topic of Catcher ERA has come up from time to time both on Batter's Box and alt.sports.baseball.tor-bluejays. A few people have argued that Ken Huckaby, although a lousy hitter, should remain one of the regular catchers on the Blue Jays due to his low Catcher ERA. I claimed that the comparison wasn't relevant because the two catchers may have caught different pitchers, so the catcher who primarily caught Roy Halladay should have a better CERA than the catcher who primarily caught Scott Cassidy.

Unfortunately, nobody ever bothered to find out how each pitcher performed when matched up with each catcher. I decided to take a look at the stats, with the help of Astros Statistical Software. I examine 14 different pitchers who had at least 10 innings pitched against 2 or more catchers.

The problem with this kind of analysis is the small sample sizes involved. Even though a pitcher might have an ERA a full 200 points lower with one catcher than another this may just be random chance as a pitcher/catcher pair may have only worked together for 15 or 20 innings. The fact that the offensive performance of opposing teams varies widely will also introduce a bias. If Darrin Fletcher caught Luke Prokopec for a couple games in Detroit, and Ken Huckaby caught him for a couple of games in Texas, then Fletcher will probably look like the much better catcher. So take anything in here with a huge grain of salt.

I look at four metrics of determining pitcher performance. The first three are the standard Home Runs Per 9, Walks Per 9, and Strikeouts Per 9. The last column, DERA is DIPS ERA or Defense Independent Pitching Stats ERA, created by Voros McCracken. To learn more about DIPS ERA see this discussion at Baseball Primer. I used the formula posted by Vinay Kumar on the bottom of the page.

Now to the results:


Brandon Lyon IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 27.3 1.65 2.31 4.29 5.32
Tom Wilson 26.3 2.74 3.76 4.11 7.25


Brandon Lyon gave up a lot more homeruns and walks for Tom Wilson than he did for Darrin Fletcher.

Chris Carpenter IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 18.0 3.00 5.00 8.50 7.08
Ken Huckaby 50.3 0.72 2.86 4.83 4.17


Chris Carpenter pitched much better for Huckaby than he did for Fletcher. Carpenter gave up more walks and homeruns when Fletcher was catching but he also struck out more batters.

Cliff Politte IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 10.0 1.80 3.60 8.10 5.22
Ken Huckaby 32.7 0.55 4.13 10.18 3.23
Tom Wilson 14.0 0.64 0.00 6.43 2.74


Politte was mediocre for the 10 innings Fletcher caught him, but was terrific for Huckaby and Wilson. In 14 innings he did not walk a single batter when Wilson was catching, but he had a much higher strikeout rate when Huckaby was behind the dish.

Corey Thurman IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 15.3 0.59 7.65 5.29 5.62
Ken Huckaby 25.0 3.24 5.40 7.56 7.93
Tom Wilson 22.7 0.40 6.74 9.12 4.02


Corey Thurman was incredibly prone to the long ball when Huckaby was catching, who he had a 7.93 DERA with. Unlike Politte, Thurman had a much higher strikeout rate with Wilson than Huckaby.

Esteban Loaiza IP HR/9 BB/9 K/9 DERA

Ken Huckaby 98.3 1.01 1.92 5.22 4.19
Tom Wilson 45.0 1.40 2.60 5.00 4.88


Esteban Loaiza gave up fewer homeruns and walks and had more strikeouts for Huckaby than Wilson, with a difference in DERA of 0.79.

Felix Heredia IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 15.7 0.00 4.01 7.45 3.02
Ken Huckaby 14.7 0.61 1.84 4.29 3.87
Tom Wilson 20.0 1.80 6.75 4.05 6.77


Felix Heredia pitched well for Fletcher and Huckaby, but was awful when pitching to Wilson.

Justin Miller IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 13.3 0.68 8.80 5.41 5.61
Ken Huckaby 71.3 1.26 5.43 7.07 5.24
Tom Wilson 17.7 0.51 5.08 2.03 5.48


Justin Miller performed about the same for each catcher who was assigned to him. His strikeout rate was much higher with Huckaby than Wilson, but unfortunately so was his home run rate.

Kelvim Escobar IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 12.3 0.00 6.59 10.24 3.25
Ken Huckaby 38.0 1.18 4.74 10.18 4.27
Tom Wilson 25.7 1.75 4.90 9.11 5.16


Escobar was much better pitching to Fletcher than he was to Huckaby or Wilson. Unlike Miller, Escobar gave up far more homeruns when Wilson was catching than Huckaby.

Luke Prokopec IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 25.7 1.05 3.85 5.95 4.66
Ken Huckaby 20.7 3.04 1.30 5.22 6.85
Tom Wilson 25.3 3.20 3.91 4.27 7.85


Prokopec also pitched much better for Fletcher than he did for the other two catchers. His control seemed to elude him when he pitched for Wilson, as he had a much higher walk rate and a lower strikeout rate for Wilson when compared to Huckaby.

Pete Walker IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 14.0 0.64 1.93 6.43 3.47
Ken Huckaby 62.0 1.31 3.63 4.94 5.21
Tom Wilson 56.3 1.12 3.36 5.28 4.79


Walker was better in all three categories when pitching to Wilson than he was pitching to Huckaby leading to a difference of 0.42 in DERA.

Roy Halladay IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 33.7 0.80 1.60 7.21 3.41
Ken Huckaby 149.7 0.36 2.40 5.53 3.43
Tom Wilson 56.0 0.16 2.57 7.88 2.73


This absolutely floored me. Roy Halladay was a much better pitcher when matched up with Tom Wilson, as his homerun rate was much lower when Wilson was catching. Halladay still pitched very well for Huckaby, who caught Halladay almost 100 more innings than Wilson did. I believe this is why Ken Huckaby had a much lower CERA than Tom Wilson in 2002.

Scott Cassidy IP HR/9 BB/9 K/9 DERA

Ken Huckaby 28.0 1.61 3.21 6.75 5.13
Tom Wilson 24.3 2.22 5.93 5.56 7.47


Scott Cassidy was a much better pitcher when Huckaby was behind the plate rather than Wilson.

Scott Eyre IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 21.3 1.27 6.34 6.34 5.44
Ken Huckaby 15.0 0.60 3.00 7.80 3.45
Tom Wilson 27.0 0.00 3.00 7.67 2.70


On the other hand, Scott Eyre did not perform well with Huckaby as a catcher and was even worse with Fletcher. Eyre did not give up a homerun in the 27 innings Wilson caught him, giving Eyre a much better DERA with Wilson as compared to Huckaby.

Steve Parris IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 16.7 1.08 4.31 3.77 5.27
Ken Huckaby 44.7 1.41 4.23 6.04 5.26
Tom Wilson 14.0 2.57 3.86 7.07 6.11


Steve Parris did not pitch well when matched up with any catcher, but pitched worst when matched up with Wilson. Parris gave up 4 homeruns in 14 innings when Wilson was catching.



That's all fourteen pitchers. To see how the catchers performed as a whole, I used the "Matched Innings" method from The Diamond Appraised. I have no reason to recommend this method of averaging over any other one, but it's probably the best known so I decided to use it. Using Wright's methodology, there are 163 matched innings over 10 of the 14 pitchers:



Matched Innings IP HR/9 BB/9 K/9 DERA

Darrin Fletcher 163.0 0.78 4.45 6.59 4.37
Ken Huckaby 163.0 1.34 3.34 6.53 4.84
Tom Wilson 163.0 1.17 3.95 6.40 4.74


Matching innings this way leads to Huckaby having the highest Catcher DERA on the team. I'm quite certain that a different method of matching would show Wilson having the highest Catcher DERA. Given the sample sizes, it's pretty difficult to determine anything from this. From what little data we have, I do not see any indications that Huckaby handled the pitching staff any better than Wilson did.

I'd love to hear your comments on the matter.
Do Catchers Matter? | 13 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.
Mike D - Wednesday, June 11 2003 @ 07:58 PM EDT (#100338) #
If there's any way to adjust for opposition and ballpark, I'd be interested in the results. Interesting starting point, though.
_Brent - Wednesday, June 11 2003 @ 08:10 PM EDT (#100339) #
Agreed. I believe that the two factors that Mike D is talking about might be confounding this study. Otherwise, this is a great study Mike.

However, if the data is indeed significant, and Huckaby did have the highest DERA, then it leads to quite an amazing fact: Huckaby made it through almost an entire year by contributing nothing above average. It seems Huckaby's saving grace was his superior game calling skills and now that might turn out to be incorrect. Interesting.
Craig B - Wednesday, June 11 2003 @ 08:59 PM EDT (#100340) #
It should be noted that a difference of half a run of ERA over 163 innings is nothing; well within the regular variance of ERA over that time.
Pepper Moffatt - Wednesday, June 11 2003 @ 09:52 PM EDT (#100341) #
http://economics.about.com
However, if the data is indeed significant, and Huckaby did have the highest DERA..

I haven't run any sort of significance tests, but I'm absolutely certain none of the results would be statistically significant. This supports the hypothesis that if there is a difference between the ability of catchers to influence a pitcher's performance, it's rather small.

It seems Huckaby's saving grace was his superior game calling skills and now that might turn out to be incorrect.

One point in Huckaby's favor: If catchers do have an influence on performance, it probably comes from their ability to "frame" pitches. If a catcher is highly skilled in that area, he should have a higher K/9 rate and a lower BB/9 rate. Huckaby did have a lower BB/9 rate, though his K/9 rate was about the same. Again, the difference isn't really that significant given the sample size.

If there's any way to adjust for opposition and ballpark, I'd be interested in the results. Interesting starting point, though.

I thought about doing that, but gave up when I realize that it would be a lot of work and might not add much.

What you would need to do is augment the K/9, BB/9, and HR/9 figures for ballpark and opponent. There are general factors you could use to do this. I don't think it would add much, though, as ballpark/opponent factors are just averages and could vary widely from game to game. This doesn't matter so much when you're using factors over 81 home games, but when you're talking about 1 or 2 appearances, you might just end up adding more noise than you take out. Wrigley in spring doesn't play like Wrigley in summer, and the Skydome is a different ballpark when the dome is closed... stuff like that.

As well, instead of using a general team factor, I think you'd have to consider each individual batter. The Brewers are a much different team when Sexson and Jenkins are in the lineup.. so the adjustment of HR/9 appearance against the Brew Crew should be a lot different if those two are in the lineup. Again, when you're aggregating over 162 games, it's not a big deal, but when it's one or two appearances, it matters a lot more.

The problem with an analysis like this is sample size issues. There seems no great way of getting around them.

Mike
robertdudek - Thursday, June 12 2003 @ 12:08 AM EDT (#100342) #
It would be interesting to select teams that had the same two catchers for a period of 5 years or more and see how they did over a larger sample. I suspect there aren't going to be a lot of those, however.
_Spicol - Thursday, June 12 2003 @ 10:26 AM EDT (#100343) #
I'll say the same thing I said to Woolner after he published his study:

There is a huge problem with the raw data used in all catcher ERA analysis. Unless you can determine which pitches are called by the catcher, which by the pitcher and which by the bench, we can't isolate and measure what the catcher's contribution truly is.

That said, a difference of 2 runs in CERA, which is the difference between Tom Wilson and Greg Myers this season and was discussed by yours truly a few weeks ago, is such a large gap that I'm confident that there is something to it even if I can't measure what it is.
Pepper Moffatt - Thursday, June 12 2003 @ 11:48 AM EDT (#100344) #
http://economics.about.com
There is a huge problem with the raw data used in all catcher ERA analysis. Unless you can determine which pitches are called by the catcher, which by the pitcher and which by the bench, we can't isolate and measure what the catcher's contribution truly is.

That's assuming most of the catcher's value comes from calling pitches. You also have to consider the "skill" of framing pitches.

As well, in the end it shouldn't really matter who chose the pitch.. the pitcher or the catcher. All that should matter is performance.

An example: Imagine there's a catcher (Catcher A) who tells his pitchers "Call your own game.. don't take my first choice of pitch as a recommendation.. I'm just going to put my fingers down in a random order". Now suppose the pitchers have a lot of success calling their own game and Catcher A ends up having a low ERA. That low ERA would indicate that the catcher is bad at choosing pitches, not good.

That said, a difference of 2 runs in CERA, which is the difference between Tom Wilson and Greg Myers this season and was discussed by yours truly a few weeks ago, is such a large gap that I'm confident that there is something to it even if I can't measure what it is.

Again, it depends which catcher is catching which pitcher. I don't think you can conclude anything from it until you look at the data closer. Unfortunately I don't have 2003 data, so I won't be churning out those figures.

Mike
_Mick - Thursday, June 12 2003 @ 12:30 PM EDT (#100345) #
There y'all go again, trying to confuse us with "facts" and "statistics." Zombies.

As for the headline, "Do Catchers Matter?," I am reminded of Casey's famous line, "You need a catcher or else you'll have a lot of passed balls."
_Spicol - Thursday, June 12 2003 @ 12:43 PM EDT (#100346) #
That's assuming most of the catcher's value comes from calling pitches.

It's assuming nothing. My assertion is that those pitches called by the bench or pitcher skews your data since you are using the results of all pitches, no matter who calls them. The catcher value equation includes many items, including pitch framing as you mentioned. I realize pitch calling is only part of that bigger equation but by including pitches called by either the pitcher or the bench, you're throwing the whole analysis out of whack.

Again, it depends which catcher is catching which pitcher. I don't think you can conclude anything from it until you look at the data closer.

I'm not concluding, though. I'm using the huge gap to hypothesize. ;) I wouldn't hypothesize if the gap were smaller...as Craig stated, there is an acceptable and regular variance. Looking at the 2003 data closer will result in the same weak conclusions that all catcher studies result in. But when one is in a leadership position, as Tosca and JP are, one has to go with scraps of information and piece them together until a picture hazily resembles something.
robertdudek - Thursday, June 12 2003 @ 01:03 PM EDT (#100347) #
Spicol,

What are the precise CERA numbers to date in 2003?
Craig B - Thursday, June 12 2003 @ 02:56 PM EDT (#100348) #
Myers's CERA is 4.32 in 295 defensive innings.
Wilson's CERA is 6.08 in 257 defensive innings.
_Jurgen - Thursday, June 12 2003 @ 04:53 PM EDT (#100349) #
Sorry, Mike, but it reminds me of a scene from everyone's favourite show:

Lisa: Hah! Here's Dr. Gould now. What were the results, professor?
Dr. Stephen Jay Gould: Inconclusive.
Lisa: Inconclusive?! Then why did you run up like that!
Dr. Gould: Can I use your bathroom?


I agree that it would be interesting to broaden the study.
Pepper Moffatt - Thursday, June 12 2003 @ 04:57 PM EDT (#100350) #
http://economics.about.com
You don't have to apologize if you're comparing me to Stephen J. Gould. :)

I agree that it would be interesting to broaden the study.

Maybe I should apply to JP for a grant.

I'd first like to look at Myers and Wilson this year, but I can't figure out an easy way to get the data.

Mike
Do Catchers Matter? | 13 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.