Quantifying Catcher's Defense

[Update, 2:38 PM] - In an email with Lee, my memory was jogged that I didn't do two things: First, I forgot to thank Lee. He was a big help in this. He likes sabermetrics, I like sabermetrics and he's got a book about sabermetrics. Go buy it. Secondly, if you'd like to do the same kind of study -- and who wouldn't? -- basically this is what I did:

This is for 2010 and I got it from here. I get names, teams, league, innings from the Standard Fielding table. I use PA's for the catchers rather than IP since it's better (a catcher must be out there for a 10 PA inning, but it would only go down as 1 inning) and I can get that plus the Fielding Erros, Throwing Errors and Catching errors from the Advanced Fielding table. Then using the Baserunning table, I get passed balls/wild pitches, stolen bases and I use CSctch which is caught stealings from catchers only -- no pick offs from pitchers.

Defense in baseball is remaining a hot topic. There's just not enough information out there that is good enough to give us results that come close to the accuracy we have regarding offensive statistics in baseball. But it makes sense, given we've tracked offensive statistics for over 100 years and defense has been tracked for a fraction of that.

But under the defense umbrella is the toughest of which to quantify: catcher's defense. There are a couple different ways to do it. There's a With or Without You (WOWY) method by Tom Tango which is just as the name suggests: how does the defense fair with and without your talents. Dan Turkenkopf and Brian Cartwright have also done WOWY work with catcher's. Unfortunately I lack the necessary database skills (read: I possess zero database skills) to pull off a WOWY study.

Fortunately,  Matt Klaassen's laid out a method that is less intrusive and still yields decent results. This is the same method that Justin Inaz did for his in-house Wins Above Replacement calculations a couple years back before Fangraphs implemented them. I will be using the method that Matt used in October of last year. There are a couple of articles that uses a 'reputation runs' number as well. I'm not entirely comfortable with that idea. The theory behind it is it gives a catcher like Yadier Molina a boost because people don't even attempt to steal against him compared to someone like Josh Bard. I'm not sure how necessary that is.

Nonetheless, you can see that this isn't something new or even ground breaking. I've talked this out with Matt and got everything as close to his results as I can. So lets get to the results; who was the best defensive catcher in 2009?

Okay, so we're not quite to the results portion of this yet. Basically, I want to run down the method I used first. What I did was find the league average in the following categories: stolen bases allowed, caught stealings, wild pitches and passed balls allowed, catching and throwing errors. Each of these has a run value attached to it (using linear weights if you want to be nerdy). The actual formulas I use are the same as in the Matt Klaassen article I linked to earlier. Click over there to get them exactly if that's your cup of tea.

Once I have the league rates, I then compare each individual catcher to the league rates, convert to a runs total for each of the above categories. You then sum those and voila! You have a catchers runs saved (or allowed) above (or below) average.

So, without further ado, your top 25 catchers in 2009, sorted by highest runs saved total.

Key for the non-obvious acronyms:

CSRns = Caught Stealing runs
WP/PBRns = Wild Pitch/Passed Ball's runs
FERns = Fielding Runs
TERns = Throwing Runs
TotalRns = Total Runs

Before we get into the Tigers aspect of this, I will say that I was relatively good in replicating Matt's numbers. Neither of us could figure out why, but I was anywhere from 0.0 to 0.5 runs off from his calculations. That margin of error isn't palatable, but isn't large enough to give me big cause for concern.

Now this leaderboard is really good news for us Detroit Tigers fans. Gerald Laird was surreal last year behind the plate defensively. The bulk of it, obviously, came on the back of his great 40% caught stealing rate. You can pile up runs saved with just your arm when you gun down 4 out of every 10 stolen bases attempts and you have the fifth most steals attempted against you in the major leagues (99 attempts). I don't know why people kept running on Laird -- maybe it was because they thought non-Justin Verlander pitchers were too slow to the plate -- but whatever the reason was: thank you opposing teams.

What does this mean? Well, it means that if you meander over to Fangraphs and take a look at Gerald Laird's Value section of his page, you will see he was worth 0.8 Wins Above Replacement. I still need to pen a Saber 101 piece about WAR (and will do so at some point), but that is essentially saying he was 0.8 wins better than any freely available talent in Triple-A.

Unfortunately, Fangraphs doesn't include any sort of defensive estimation for catchers. Why? I don't know. I calculated data for the Fangraphs WAR Era which is 2002-present day in about 2-3 hours of actual work -- most of that just copy and pasting from Baseball Reference. So I find it a bit inexcusable that they would leave out the bits that can be measured about catchers defense, but I digress.

When you look at Laird's minuscule 0.8 WAR, you add on another 1.37 wins to that. He was actually worth 2.17 WIns Above Replacement in 2009 before you get to base running (another downfall of the WAR calculations at Fangraphs; Laird was a positive base runner last year, as well).

Other people of note:

• Like Joe Mauer needed more value added to him, you can tack on another 0.5 wins to his 2009 WAR total.
• David Ross did a lot of damage in a small (relatively speaking) amount of innings defensively for the Braves.
• Toronto featured some nice backstops last year in Rod Barajas and Raul Chavez (both 4.6 runs saved). And they had Greg Zaun, too, who combined between Baltimore and Toronto put up a positive number of runs saved.
• The spread between the best (Laird) and the worst (Mike Napoli) is 23.1 runs. That is basically 2-wins worth of value between what the Tigers had starting behind the plate and what the Angels had.
• Pudge Rodriguez has some defensive value, still.

There are caveats here, like with nearly any defensive measure. I don't think there's a way to quantify the impact a catcher can have in terms of game calling abilities. Also, the catcher's defensive abilities are somewhat tied to the pitcher. It's hard to throw people out if your pitcher takes the longest time to get the ball to home plate with a runner on. Kenny Rogers is also going keep a guy closer to the bag than, say, Jeremy Bonderman would. This doesn't happen really with any other position on the diamond. Additionally, the type of pitcher you're catching affects how good you are. Josh Bard spent a large chunk of time catching Tim Wakefield. This type of study can't/doesn't adjust for that. Keep in mind the catcher actually fields the ball so little relative to the other 7 defenders, that it's probably the most unstable position to get a defensive read on.

As with all defensive measure, the larger the sample, the better. One year is not enough to make conclusions on, and that's why I've collected 8 years worth of data. Gerald Laird is probably not going to save nearly 1.5 Wins worth of runs for the Tigers again -- there is great amounts of regression for a player year-to-year.

That said, something is better than nothing. I plan to have updates at the end of each month for defensive runs saved for catchers, and I have the data that goes back to 2002 like I said. My next post in this series will look at the best and the worst over the last 8 years behind the plate.