Thursday, September 03, 2009
How much do you really value defense?Posted by Pat Andriola
Here's the situation: you're the GM of a well-balanced team heading into the 2008 baseball season. You have adequate pitching, defense, and hitting, and all you are now looking to do is round out your team by acquiring a right fielder. Two are available for you, and we'll assume that both are the same age, will cost the same price monetarily, and have the same durability. Also, by the way of some awesome forecaster you are 100% sure that this will be their offensive numbers for the upcoming season:
Player A: .283/.381/.498, .379 wOBA
Player B: .249/.323/.406, .321 wOBA
Now, here are their defensive numbers for the upcoming season, again with 100% certainty:
Player A: -46.6 UZR/150
Player B: -10.5 UZR/150
Finally, here are their respective WAR totals for the year, again with certainty, and this time revealing their identities:
Brad Hawpe: -.7 WAR
Jeremy Hermida: .1 WAR
To stick with the situation we were given, we are really only concerned with finding the best right fielder possible for the 2008 season (let's forget long-term goals). To sum it up, Hawpe is good offensively, but is in another league in terms of his miserable defense, which gave him negative value. Hermida is pedestrian offensively, and has pretty bad defense as well, but does just enough on both ends to give him at least some positive value.
So the question you now have is: how much do you really value defense? Let's remember that UZR is not a rate stat; it is specifically telling us that Brad Hawpe will cost his team -46.6 runs (~4.7 wins) thanks to his defense, whereas Hermida will cost his team -10.5 runs (~1 win) thanks to his. Can we fudge these numbers? Maybe, but probably not by much. MGL gives us two types of measurement errors, but for our purposes only the second one applies:
In the case of defensive metrics we have large and numerous measurement error. For example, that one ball that Yuni got to that is only gotten to 20% of the time by the average fielder, may NOT have had the characteristics we thought it did. While all hard hit ground balls hit in area X up the middle may be fielded 20% of the time, there are obviously lots of different types of balls in that bucket, from an easy, high chopper (albeit hard hit, if that is our bucket) to a ground skinner. As well, some of those balls in that bucket are hit directly up the middle and some are hit a little to the SS or 2B side. In addition to that, sometimes the SS is playing in position A and sometimes in position B because of the batter, pitcher, base runners and outs. Lot and lots of potential measurement error. So that ball that Yuni fielded may have been an easy chopper a little towards the SS side, even though it went into the “hard hit ball up the middle” bucket which is only fielded 20% of the time on the average by all SS.
I think this question is ultimately very similar to the post Dave Cameron made at Fangraphs when he compared Adam Dunn to Nyjer Morgan. However, the difference here is that Hermida isn't some defensive superstar; he's actually very bad himself. But when we're talking about a significant difference in offensive value, how much weight do we put into the defensive difference when both are bad themselves? I ultimately think that it doesn't really matter if both are "good" or "bad;" we have the data, and the data clearly shows us just how much better/worse one player is than the other. However, John Dewan, of "The Fielding Bible" and Plus/Minus fame, recently said that:
On offense I believe we’re measuring 80-90 percent of the true ability of players. On defense, I believe we’re at about the 60 percent level. But we’re still at the tip of the iceberg in terms of precision and a ton more can be done, especially defensively. As new forms of data become available, we’ll be able to enhance our defensive systems. One example: BIS has now developed a batted ball timer, which we believe will greatly improve the accuracy of our system.
If we're going to take this at face value, we then may be more inclined to take Hawpe, as we are more confident in accuracy of the offensive stats compared to the defensive ones. But then again, the difference between Hawpe and Hermida's defense was so huge than even making some changes based on possible measurement error/inherent problems with the system might not be enough to put Hawpe over the top.
I think Rob Neyer summed up the frustration/hope involved in defensive metrics best when he said, "Maybe we don't completely get the value of defense yet. But trust me, Dave Cameron: we're getting there, and we're getting there quickly." I hope (and think) that he's right, because I'm still not fully sold on which player I'd take. I'm leaning towards Hermida (barely), but I'd still like to hear from others.