Catching Sabathia: Jorge’s vs Jose’s way

Today we are going to compare C.C. Sabathia’s pitching patterns when he is caught by Jorge Posada with those when he is caught by Jose Molina.

I know, I know. It’s A.J. Burnett who wasn’t in the same page with Jorge, prompting manager Joe Girardi to team him with Molina for the final part of the season, playoffs included.

However, Burnett’s starts with Posada happened in a different moment of the year than his starts with Molina, so any change in Burnett’s behavior/effectiveness, while probably due to his battery mate, might also have had something to do with his own evolution during the season.

We’ll never see a random pattern in pitcher-catcher teaming, so a distribution of starts like Sabathia’s (next chart) isn’t too bad for an exploratory analysis.

image
Figure 1. - Catchers in C.C. Sabathia's starts (regular season).

Starts with Posada and starts with Molina are fairly well intertwined, except for a long span from start No. 7 seven to start No. 18 when Molina never caught Sabathia. Furthermore, we have nearly 400 at-bats with Posada catching and nearly 300 with Molina catching, plus another 250-something with Francisco Cervelli behind the plate.

The stuff

According to scouting report at 60ft 6in , Sabathia’s repertoire consists of a “a great fastball from the left side, a slider and a changeup”; the slider is said to move “like a hybrid slider/curve type pitch.” Fangraphs data (2009 figures) show him throwing the heater around 62 percent of the time at an average speed of 94 mph and going to the slider and the change of pace at a 20 and 18 percent rate respectively.

However, with PITCHf/x data and a little help from cluster analysis and Matt Lentzner’s Peanut chart, I re-classified Sabathia pitches in four groups, according to their speed and movement.

image
Figure 2. - Classification of Sabathia's pitches, by speed and movement.

He actually throws two kind of fastballs: a four-seamer and a two-seamer (we were treated to a number of close-ups of the latter by FOX during the postseason).

Selection

Let’s see how the catchers call Sabathia’s game.

image
Table 1. - Percentage of pitches thrown to each catcher.

Molina seems different from Posada (and from Cervelli too). He relies more heavily on the sinker both against left-handed batters and against righties. He also trades, against righty hitters, a small (around three percent) portion of the change-ups with an equal quantity of sliders.

Working the count.

Here’s the game-calling according to the ball-strikes count.

image
Table 2. - Percentage of pitches thrown by count (left-handed batters).

If you swing from the left side and you’re behind 0-1, be prepared for the slider if Molina is catching Sabathia. You’ll hardly see the pitch on a 2-0 count. Other than that, the difference between the top two Yankees receivers consists of Molina’s greater propensity to call for the sinker.

image
Table 3. - Percentage of pitches thrown by count (right-handed batters).

For righties, the clue is that Molina is more reluctant (on the order of 10 percentage points) than his colleague to put down the sign for the change-up on counts 0-1, 0-2, 1-0 or 1-1, when he prefers to go with one of the fastballs. On 2-0 and 3-1 the reverse is true. Behind 0-1 and 1-2 you’re more likely (around eight percent points) to see a slider if Molina is squatting behind you.

Location

The location of each of Sabathia’s pitch types doesn’t vary with the catcher on duty, thus I won’t show any table.

Posada, Molina and Cervelli all expect Sabathia’s fastball in the same spots: outside, and preferably low, against lefties; middle-to-high against righties.
The sinker is usually requested and delivered on Sabathia’s throwing arm side (i.e., inside for lefties, outside for righties). The slider’s preferred spots are down and in against righties, and low and away against lefties. Finally the change-up, a weapon used only against right-handed batters, is aimed regularly in the low-and-outside quadrant.

Outcomes

The different slider/change-up pattern we outlined a couple of paragraphs ago results in a higher percentage of whiffs for right-handed batters against Molina, on both pitches (40 vs 36 percent on change-ups, 34 vs 25 percent on sliders). In the case of the change-up, the higher number of swings-and-misses is partly due to the greater propensity of batters to chase pitches out of the zone (56 percent of swings on change-ups out of the zone against Molina, 34 percent against Posada). The numbers for left-handed batters are too small to perform similar comparisons in a meaningful way.

A Hardball Times Update
Goodbye for now.

As you see in the table below, it seems that Molina’s game-calling produces better results both against right-handed and left-handed batters.

image
Table 4. - Pitch run value by catcher.

Looking at the run values split by pitch type, we see that Posada’s sequences against lefties rely in great part on the fastball effectiveness, while Molina gets the most out of Sabathia’s sinker and slider. Against righties, both catchers get the best run value out of the change-up, but again, Molina’s mixing seem to yield better results (go back to “Working the count” to see how Molina favors the change-up in different counts than Posada).

NOTE: According to what we saw last time, the pitch run values don’t imply that Sabathia throws better change-ups to Molina, only that those pitches might be better leveraged the way they are mixed.

Now, it looks like Molina’s game-calling is more effective than Posada’s, but I invite you not to jump to conclusions. First, it must be noted that Sabathia went on a roll in the second half of the season and, while the two catchers pretty much split duties during that part of the year, Molina caught him only a couple of times when the pitcher was less than stellar early in the season.

Also, Molina’s reliance on the sinker came only in the second part of the season. Maybe, when he caught Sabathia back in May, he didn’t want to call a game too different from what the pitcher was accustomed to; maybe Sabathia grew more confident in the two-seamer as the season progressed, thus prompting the catcher to call more of that pitch. Anyway, Posada didn’t seem to follow suit—until the World Series.

image
Figure 3. - Percentage of sinkers by start (catchers: P=Posada, M=Molina, C=Cervelli, K=Cash).

Who knows what happened when the Yankees and Phillies crossed swords? Were the Phils a team to be treated with sinkers? Did Sabathia spur Posada to call the game Molina’s style?

Summing up

We have seen that Molina’s game-calling differs from Posada’s, at least when they catch Sabathia’s pitches. Unfortunately, evaluating the effectiveness of each approach is a really difficult task. It’s impossible to separate the effect of the catcher from what can be a natural evolution of a pitcher during the year.
Furthermore, even when around 300 at-bats are available for each tandem, when digging into details (and splitting numbers across many variables) sooner than later we clash against sample size issues. One more time, the run value of a pitch is heavily influenced by the ball-strikes count.

We have left aside another important, but even more difficult to evaluate, issue: Catchers may improve the effectiveness of a pitch by giving a better target, or by framing the pitch better, thus inducing more borderline calls by the umpires. For example, Cervelli’s results (not shown in this article) are in line with Molina’s, though he caught Sabathia exclusively in the first half of the season and his game-calling resembled Posada’s a lot.

References & Resources
Matt Lentzner- PITCHf/x and Pitching Models. Presented at the 2009 PITCHf/x Summit, downloadable here.

Yankees scouting reports at 60ft 6in, by Sven Jenkins.

PITCHf/x data from MLBAM; pitch classification by the author.


9 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Peter Jensen
14 years ago

Another great article Max.  Have you looked into whether there are any differences in the quality of the opposition batters that the different catchers faced?  It would seem to be important to control for that.  It might also be interesting to see if the pitch mix differences for above average hitters and below average hitters of each handedness.  It may be that the catchers call for similar pitches for one and not the other.

Dave Studeman
14 years ago

Max, this is beyond awesome.  I particularly like the format of the graphs and tables.  Very readable stuff.

Sean Smith
14 years ago

This is good stuff.  But I think seeing what AJ Burnett throws to different catchers would be more interesting.  Maybe throw in what he threw to Toronto catchers in 2008 as well.

Nick J
14 years ago

Thank you.  This is great.  I second Studes’ compliments on the graphs as well.

Benny G
14 years ago

Stat from the best baseball book I have read this off-season: Hardball Times 2010.  The yankee team ERA was over 5 when Posada caught and 3.5 when he didn’t.

Ray
14 years ago

Benny..I don’t have the numbers to delve into here but its possible that Posada always catches the fill-in starters of the staff, and sometimes the aces, where as Molina gets to play only with the good pitchers and never any of the bad ones.

OldYanksFan
14 years ago

Also, if you are judging the ‘results’ by Runs, don’t ya have to look at the BA/OPS of the batter being thrown to? Does Posada Catch more against better teams?

This is great stuff, but as mentioned, there are many variables that effect the outcome. It’s hard to make too many judgements without some of those variables being resolved.

However, to simplify it, it looks like Posada favors the fastball. Does he have problems catching pitches with movement? Is he afraid of the stolen base? There were rumors that Pudge called too many fastballs with men on base, so he had a better shot throwing them out.

Is there a reason Posada seems to favor the FB?

HCrumb
14 years ago

This analysis is interesting technically (e.g. use pitch f/x) , but it’s statistical methods aren’t particularly strong. 

Minimal attempt has been made to control for factors that likely confound and/or bias the results such as the level of competition each pitcher/ catcher was facing.

Neither catcher was randomly allocated to the pitcher.  As such, in order for this analysis to produce valid findings, extensive conceptual and methodological work would be needed to to address bias. In addition, use of quasi-experimental analytic methods would have greatly improve the validity of inference that can be drawn from the data.  Even then only very tempered conclusions could be drawn.

Conceptually, it would have been important to assess what the inherent validity of the underlying hypothesis being evaluated is as well.

Phil
14 years ago

Very nice article.  I was surprised to see little mention of Posada’s and Molina’s respective abilities to block pitches in the dirt.  It seems that Molina’s reputation as the better defensive catcher would support greater comfort calling for sinkers (particularly with men on base).