Two dimensions of catching—and dealing with interactions.

A catcher’s defensive game has many facets.

Much work has been done to quantify the value they contribute by controlling the running game and by avoiding passed balls and wild pitches; something also has been tried to evaluate other elusive aspects of catchers’ fielding, namely framing pitches and calling the game.

Today I’m going to add my footprints to the more walked path, though I know I would make many more fans if I showed a way to measure game-calling ability.

The main issue in evaluating a catcher’s ability in preventing balls from rolling toward the backstop is that pitchers have a great influence on the outcome: You can not forget that Geno Petralli caught (or at least tried to catch) many of Charlie Hough‘s knuckleballs when you look at his passed balls and wild pitches allowed totals in 1987. Tom Tango explained the matter really well in an article published in The 2008 Hardball Times Annual.

When considering baserunning, you still have to take pitchers into account: Some take forever delivering the ball, while on others you haven’t a chance to get a decent jump. Then, obviously, there is the baserunner; thus the catcher is just one of three actors in the running game.

Tango proposed the WOWY (With Or Without You) method as a way to separate the effect of two actors on the same play. Again, I point you to the 2008 Hardball TImes Annual article (actually there are two of them) for reference. If you didn’t buy the book a couple of years ago, there is also this introduction on his website (Evaluating Catchers @

Tom himself outlines a potential problem of WOWY, the Cal Ripken effect as he calls it. Some players have a very small sample in the “without” bucket: in the pitcher-catcher interaction it can happen when one receiver becomes the personal batterymate of a specific hurler. In an interaction between pitcher and shortstop it has happened to whatever pitcher played only for the O’s in the ’80s: he always had Cal at short (hence the name for the effect).

In such cases it becomes really hard to disentangle the effect of each player in the action, since comparison with other matchups is not available.

It is beyond the scope of this article to introduce an advanced statistical technique (have a look at the Resource and References section); suffice it to say that, in this case, it is a good (likely better) alternative to the WOWY method, and that through simulation it has been proved that it gives good results even when the combinations of cross-classified levels (pitcher-catcher diads in our case) have very small numbers of observations.

Passed balls and wild pitches

For each diad, I found—via the invaluable Retrosheet—the number of plate appearances with runner(s) on base (i.e., the potential occasions for wild pitches or passed balls) and the sum of wild pitches and passed balls occurred. Then I calculated the percentage of “missed balls” as the ratio between the two numbers.

(I assigned a rough 0.3 run value to each missed ball, just to have a very rudimental translation of the percentages to a run value. This is an area of my research that needs to be worked on in the future.)

The multilevel analysis spits out how much the presence of a catcher behind the plate (or a pitcher on the hill) influences the likelihood of missed balls. The results show a high degree of correlation with those from Tango’s WOWY method. The Pearson’s correlation coefficient between the two methods is .81 (95 percent Confidence Intervals: 0.76 – 0.86).

Looking through the lists I find Yogi Berra, Brad Ausmus and Gary Carter among catchers historically good at preventing balls from rolling toward the backstop (Yogi is actually at the very top, followed by Bruce Benedict and Jake Gibbs), and Sandy Martinez, Mike Rivera and Bob Uecker (way to complement his .200 lifetime batting average!) among receivers who likely had other assets in their game (which ones for the Ueck?).

On the pitchers’ side I find Hoyt Wilhelm, Charlie Hough and Scott Williamson as the wildest ones (all the noted knuckleballers are in the discussion) and Greg Maddux, Luis Tiant and Brad Radke as the easiest to catch.

(Here are my full lists, available as comma-separated text files (pitcherscatchers). I invite you to look at them and express whatever concern you have about the rankings. As I stated before, at this stage my work on translating to run value has been very poor: the r100 column should be interpreted as something like “runs saved per 100 plate appearances with men on base.”)

One issue of WOWY that still remains with multilevel modeling is that a strong assumption is being made on the pitchers’/catchers’ abilities over time. Basically we are assuming that wildness and blocking ability remain constant through the players’ careers. Let’s look at Randy Johnson‘s wild pitch career trend.

Mental Health and the CBA
A particular bit of language in the latest CBA could have negative consequences for some players.

(I cheated a bit on this one, removing Johnson’s last season, which saw him throw a high number of wild pitches, thus making the smooth line go up to mid-’90s heights.)

Clearly, catchers who were teamed with him in his Mariners days get penalized in comparison to his Yankees and D-backs receivers. Similarly a pitcher might get worse numbers if, by (hard) luck, he always played with catchers at the end of their careers, when bad knees had eroded the skill of going down to block pitches in the dirt.

The running game.

Here the multilevel approach gives us another important advantage over WOWY: we can put three actors into the mix, without too much effort. This time I limited the field to situations with runner on first (only) and assigned (to each pitcher/catcher/baserunner triplet) a +0.3 value every time second base was swiped and -0.6 for each runner gunned down.

The correlation is lower this time between multilevel and WOWY results for catchers (Persons’ rho: 0.43 ; 95 percent Confidence Intervals: 0.31 – 0.55).

This time there are three lists to skim through.
Catchers: Ivan Rodriguez, Clay Dalrymple and Steve Lake come out on top; Joe Nolan, Craig Biggio and Alan Ashby close the list.
Pitchers: Pete Vuckovich, John Denny and John Candelaria helped their catchers a lot (I never would have thought Gaylord Perry is just behind those three); John Montefusco, Nolan Ryan and Steve Bedrosian didn’t care too much about baserunners.
Baserunners: Willie Wilson, Vince Coleman and Lou Brock were the best once you factor in the battery against which they were stealing (Wanna know about Rickey? Fifth, behind Tim Raines; best among the active players? Carl Crawford, barely missing the top 10); finally the worst three on the basepath, Retrosheet era, are Duane Kuiper, Minnie Minoso and Greg Gross.

Here are the files (same disclaimer as above applying on interpreting the numbers); any observation welcome: pitcherscatchersbaserunners.

Again we have to be aware of the “flat career assumption” previously outlined.

Another issue, probably relevant in the wild pitches/passed balls section, too, is how catchers (pitchers too?) alter the rest of their game to better control the baserunning. A backstopper might be so preoccupied with the runner on first that he alters the pitch selection, requesting lots of fastballs. As a result, he is probably adequate at keeping the runners honest, but his pitcher’s effectiveness is going to suffer.

Where to go?

The method seems to yield reasonable results, as it’s quite in accordance with Tom Tango’s findings. Furthermore, skimming through the lists we find expected things: Piazza adequate at blocking pitches but awful at controlling the running game, all the knuckleballers being difficult to receive (the opposite being true for the finesse pitchers), the usual suspects as the top runners.

In the future I would like to use the same approach for the pitch-framing and game-calling spheres, but I believe a solution to correct for the aging effect is needed first.

The same method can be used for other situations where an interaction between players plays a significant role in the outcome. I can name a few on the top of my mind: outfielders’ arm and baserunning, protection by the following hitter and scooping by first basemen (as has been done a couple of times using WOWY).

References & Resources
The information used here was obtained free of charge from and is copyrighted by Retrosheet. Interested parties may contact Retrosheet at 20 Sunset Rd., Newark, Del. 19711.

Tom Tango: With or Without You and With or Without… Derek Jeter both appearing on The 2008 Hardball Times Annual; plus Evaluating Catchers on his website.

Multilevel Modeling
This site is a terrific resource on multilevel modeling; in particular I would suggest this paper (PDF) on cross-classified and multiple membership structures in multilevel models.

Print This Post
Sort by:   newest | oldest | most voted
Dave Studeman
Dave Studeman

I can’t comment on the methodology, Max, but this is terrific stuff.  Are you thinking that you can create a solution to the aging issue?

Also, people can read the 2008 Annual for free here:

Michael Lahr
Michael Lahr

It’s interesting to see how low Varitek and Piazza rank here.

Tom M. Tango
Tom M. Tango

Great stuff Max.  I’d be interested to hear more about Brock v Raines.  The gap between the two is 130 SB and 161 CS.  It’s hard to believe that Brock can be in the same ballpark as Raines.

What you are proposing is that Brock ran against much tougher pitcher/catchers.


Actually, Piazza and Varitek do really well on passed balls/wp’s, but poorly on the running game.

Max Marchi
Max Marchi

Tom, looking at the SB2 and CS2 on Baseball-Reference, Raines has an 85% success rate and (roughly) 141 SB-runs, Brock 77% and 99. Thus your point is strong.
I’ll have to check the model, or the opponents against which they tried their steals.

Thanks for pointing this out – that’s why it’s useful to share the full lists.

Tom M. Tango
Tom M. Tango

Max: check out my blog.  I think we’ve got some really good insight going there…

Max Marchi
Max Marchi

Yes Tom, after I posted here I went to your blog and I thought “Wow!”.
Now I’m going back and forth.

The link, for the benefit of all is:

Northern Rebel
Northern Rebel

On the comments section of Brian Giles’ article, I’ve noted the difficulty in determining the greatness of catchers.

Who’s better Jorge Posada, or Jason Veritek?

Would the Yankees have won all those titles without Posada?

Did veritek’s leadership contribute largely to the Sawx rise?

What if they had been in each other’s place?

Are either of them HOF candidates?