Spring training is the time we hear the most about defence. Loud discussions of “getting back to basics” and “focussing on fundamentals.” Basically, all crap. But defence remains invaluable, a huge part of run prevention; itself a huge part of winning.

As fans we have access to an increasingly large number of advanced defensive statistics. Some use video technology to evaluate each batted ball against a league-wide databazzzzzzz sorry, I know it can get a little dry. The point is this: fielding percentage is dead. But there are so many current pretenders to the throne it can be hard to decide which is fit to wear the crown.

It is tough to advocate any one system over the other, in fact I like to use them all in concert. If a guy rates well in multiple systems, it is safe to say he’s decent with the leather. What really interests me when the systems disagree. Join me after the jump where I compare some of the most prominent defensive stats.

The most prominent current defensive metric is ultimate zone rating or UZR. Used in Fangraphs’ WAR calculation, UZR is a video-based system which has come under fire of late. While not perfect, it works best when you look at a large (3-year?) sample. Pack a lunch and read the full explanation at Fangraphs.

Defensive Runs Saved is also known as plus/minus and currently falls under the Bill James Umbrella of Nerditry. DRS uses video much like UZR but parses the data in a different way, as we will see below. Read a full FAQ at the Fielding Bible homepage for more insight.

Total zone is the creation of noted Sabr-dude Sean Smith. The system doesn’t use video data but play-by-play and batted ball types to divvy up responsibility among fielders. Read the full explanation at Baseball Reference, which uses TZ as the defensive component for their version of Wins Above Replacement.

Below is a three year (2008-2010) sample of qualified outfielders, comparing UZR to TZ. Why outfielders? In my mind there shouldn’t be a much grey area for batted balls once they reach the outfield. I’m trying to give the systems the benefit of the doubt, really I am.

The biggest thing to look for is disagreement between the samples. The upper left quadrant shows players who UZR doesn’t like but TZ loves while the bottom left is the reverse. This disparity fuels large differences between rWAR and fWAR, which undermines the entire process. Any names jump out? Click the image for a larger, clearer version.

Click for full sized version

Michael Bourn and Hunter Pence! This, to me, is more than mildly shocking. Defence has long been a calling card of the speedy Bourn, while Pence is very well known for his incredible throwing arm. That these two bros are teammates makes me wonder if Minute Maid Park doesn’t play a huge factor. One huge failing of play-by-play data is the perception of stringer bias; wherein one man’s line drives are another man’s bloopers. Let’s look below at the DRS and UZR comparisons to see if anyone else jumps off the screen.

Click for full version

J.D. Drew! Everything I’ve been told is a lie! Drew ranks as excellent by UZR but sits just below average when viewed by Bill James. Jeff Francoeur is always the lightning rod, here feeling some love from the DRS people while UZR isn’t quite as bullish.

Looking at the two graphs at the same time, a few names jump out as guys UZR hates but the other two players love. Torii Hunter has nearly identical DRS and TZ numbers (in the positive) but feels UZR’s scorn. Ryan Ludwick and Colby Rasmus both get a boost when you look at video versus the detailed box score.

In the end, there is a lot of agreement. Death to Flying Things, Ichiro, and Carl Crawford are good at their jobs while Brad Hawpe may or may not be paralyzed. Vernon Wells and Delmon Young make all the systems weep just like Jayson Werth and Justin Upton make them swoon.

One last bit of info: a list of the top 10 guys who confound the systems the most. The total amount of disagreement between the three systems. It is worth noting the systems all agree Jermaine Dye is awful, they simply express their hate in varying degrees.

  1. Hunter Pence.
  2. Michael Bourn
  3. Jeff Francoeur
  4. Alfonso Soriano
  5. Torii Hunter
  6. Shin-Soo Choo
  7. Jermaine Dye
  8. Ryan Ludwick
  9. Curtis Granderson
  10. B.J. Upton

Comments (21)

  1. I fully admit, these defensive calculations make my head hurt. Do any of these systems try to factor in defensive positioning?

    • Not yet, the new systems will. That is one reason you need to take a long look, over time differences in positioning tend to even out. You’re left a better sense of a player’s skill level.

  2. I dunno. Wells has definitely tailed off defensively, but is he really that level of dogshit? I think who a player’s on the field with skews these numbers somehow. Wells plays with Rios and he wins Gold Gloves, he plays with crappy range players like Joey Bats and Adam Lind, he’s useless. I’d bet if you stuck Mike Cameron — who we all know is a great defensive CF — between Lind and JB he’d have lower numbers.

    It might explain why both Bourn and Pence are rated so highly (they help each other’s numbers).

    And Carl Crawford. I guess he gets to everything, but he plays the easiest corner outfield position and can’t throw. He’s far from the best defensive outfielder in baseball, which is basically what’s said here. Then again the players who seem great when you watch them — Ichiro, Werth, Jones, Drew, all rate highly. This whole thing gives me a headache because I don’t trust it anywhere near as much as I do offensive numbers.

  3. I like how it says Ichiro! on the first graph

  4. You have make those types of adjustments on your own. Carl Crawford is great at his job (left field) but this isn’t an absolute ranking. This just shows how well they do when you compare the three numbers.

    Gold Gloves are deeply flawed, most players get them two or three years after they do their best work as word needs to filter through the league.

  5. @eric – it should really just say that everywhere his name is written. That’s how I read it regardless.

  6. @Eric You beat me to the very same observation. As Parkes said, it truly is the only way to read his name.

  7. @Theo – ahem! Parkes doesn’t spell his name like mine nor is he cursed with my good looks.

  8. I didn’t read the name of the author, but I just assumed it was Drew based on the content. When I saw “Ichiro!” I knew my assumption was confirmed.

  9. Can anyone ID the man whose silhouette appears in the photo?

  10. Drew, has there been any studies or research, that you know of, into how importantly a plus defense factors into run prevention?

    I found the beyond the box score article you linked to interesting, however I have to wonder how much of that run prevention is plus pitching, and how much is plus defense. Personally, and I have no data to base this on (*gasp* the cardinal sin!), I anticipate that the majority of run prevention is dictated by the quality of the pitching staff.

    Maybe I will find some time over the weekend to delve into this a bit.

    P.S. Sorry for the double post.

  11. Triple Post!!!!

    This is a total guess, but…. Jim Edmonds?

  12. Re: silhouete. Crawford holds his glove like that, with the webbing wide …

  13. Because of your hint, and your well documented mancrush….. Ichiro?

  14. I am afraid to guess again in case I am wrong again. There are a few ‘lantern-jawed’ people in that quadrant… Is he in that quadrant on both graphs? I am leaning towards Matt Holiday or Jason Werth.

  15. GutiĆ©rrez… that became considerably less fun with each failure.

  16. Yay! There is also the fact that a persons silhouette can vary substantially based on their orientation relative to the light source.

Leave a Reply

Your email address will not be published. Required fields are marked *