Wednesday, December 29, 2010

Catcher Defense - Part 3

Today, I'll do part 3 of my series on catcher defense.  The first two parts of the series can be found at the links below:

Catcher Defense - Part 1

Catcher Defense - Part 2

In Part 2, I looked at metrics for stopping the running game, blocking pitches and avoiding errors.  In this installment, I'll combine all of these measures into one and try to determine how many runs each catcher saves or costs his team.  My method is not a new one, but rather a variation of what others have already done.  Others include Sean Smith, Justin Inaz and Matt Klaasen.  In fact, Matt has already done this for 2010 and his numbers are very similar to mine.  I'll add that Mike Rogers has also used a similar method evaluating catchers from 2002-2009.

There are a couple reasons why I'm going to basically repeat what others have done.  First, I want to lay out the whole method for those who haven't seen it.  I'm also going to add to it in a future post, so I want it to be clear what I'm doing.

First, we need the following statistics for each catcher: Innings (Inn), stolen bases attempted (SBA), caught stealing (CS), wild pitches (WP), passed balls (PB), throwing errors (TE) and fielding errors (FE). The numbers for Victor Martinez are:

Inn = 904
SBA = 126
CS = 27
WP = 37
PB = 4
FE = 1
TE = 5

Then we need to calculate four league rates:

CS Percentage = Lg CSRate= CS/SBA = .2762
WP plus PB per inning = Lg WPPBRate = (WP + PB)/Inn = .0452
TE per inning = Lg TERate = TE/Inn = .0055
FE per inning = Lg FERate = FE/Inn = .0016

Now, we can use the above numbers as the basis of the calculation of runs cost/saved by Martinez.

We know that runners attempted to steal on Martinez 126 times in 2010.  Based  on the .2762 Lg CSRate, we would expect the average catcher to throw out 126 x .2762 = 34.8 runners in 126 opportunities.  Martinez threw out 27 runners attempting to steal, so his caught stealing rate above/below average (CS+) was -7.8.

Based on linear weights, the average caught stealing is worth 0.63 runs (0.44 for the CS plus 0.19 for the SB not achieved).  So, caught stealing runs above average (CSRuns) can be computed by multiplying CS+ by 0.63.  For example, Martinez had -7.8 x 0.63 = -4.9 CSRuns.  This means he cost his team about five runs more than what would be expected from the average catcher given the same opportunities. 

Similar calculations can be done for WP and PB.  Martinez caught 904 innings in 2010.  Based on the .0452 Lg WPPB rate, we would expect the average catcher to allow .0452 x 904 = 40.9 WP and PB in 904 innings.  Martinez allowed 41 WP and PB, so his WP plus PB above/below average (WPPB+) was +0.1.

Based on linear weights, a WP or PB costs -0.28 runs.  Thus, WP plus PB runs above/below average (WPPBRuns) is equal to WPPB+ x -0.28.  Martinez had 0.1 x -0.28 = -.03 WPPBRuns in 2010.  So, he cost his team .03 runs more than expected in preventing WP and PB.  Of course, that is essentially no runs, but I left the decimal in there in order to illustrate the calculation.

Catcher throwing error runs (TERuns) are calculated the same way as WPPBRuns.  A catcher throwing error typically occurs when a catcher attempts to either throw a runner out stealing or pick off a base runner.  Since the result is often similar to a WP or PB, we can use the same linear weight for TE as we do for WP and PB (-0.28).

Catcher fielding error runs (FERuns) are calculated the same as WPPBRuns and TERuns except that a different linear weight is used.  A catcher fielding error generally has a similar effect to errors used by other fielders (about a half run), so we can use -0.50 instead of -0.28.  Martinez had 0 TERuns and 0.2 FERuns.

(Note that the linear weights I used for errors are different from what others have used.  There doesn't seem to be a consensus and I really am not sure whether my weights are better or worse than others.  It turns out that catcher errors are so infrequent that the choice of linear weights rarely makes much of a difference in the final result)  

Finally, all of the above run values can be combined to arrive at catcher runs saved above/below average (CatchRuns):

CatchRuns = CSRuns + WPPBRuns + TERuns + FERuns

For Martinez, that is -4.9 = 0.0 + 0.0 + 0.2 = -4.7.  So, by this method, Martinez cost the Red Sox 4.7 runs more than the average catcher.

The statistics for all catchers with at least 500 innings in 2010 are shown in the table below. Cardinals catcher Yadier Molina was number one with 14.9 runs saved.  His brother Benjie was at the bottom (-13.5).  Alex Avila finished at -1.2.

The raw data for this article were abstracted from

Table 1: Catcher Runs Saved/Cost in 2010

No comments:

Post a Comment


Blog Archive


My Sabermetrics Book

My Sabermetrics Book
One of Baseball America's top ten books of 2010

Other Sabermetrics Books

Stat Counter