Wednesday, August 15, 2007

Zoned In

Last week there was a rather lengthy discussion about defense and defensive stats here on the crappy/unpopular. As a followup to that, 1060west commenter Goldwater was kind enough to put together the following post and spreadsheet concerning zone rating and how it applies to the 2007 Cubs. Below is what Goldwater found:

Gaius Marius has gone to great lengths arguing here that the Cubs have been very lucky with regard to the outcome of balls in play. I took issue with that notion, as it seemed perfectly obvious to me that the Cubs defense was simply doing a better job than most at converting batted balls into outs. While I maintain that stance, as GM has pointed out, luck and good defense are virtually indistinguishable, speaking statistically.

I thus set out to see whether we can find evidence of luck in defensive metrics. Defensive statistics have been much maligned, and rightfully so, but I think that we can glean some useful information from them, so, let’s inject some numbers into the debate.

I took a look at the zone ratings over at The Hardball Times . Essentially, these ratings

  1. assign zones of the field to each defender
  2. record the number of balls hit into those zones (BIZ) and plays made on those balls for each defender
  3. Record the number of plays made by each defender outside of his zone.

The difficulty in interpreting these ratings is that each position seems to cover its own zone at vastly different rates. In attempts to circumvent this, Justin Inaz has developed a formula for deriving the number of plays each position player makes relative to the average defender, a “+/-.“ Essentially, available data (2004-2007) is used to calculate an average rating for each position. Based on the number of balls hit to your favorite defender, we can see how many he converted to outs relative to the average at his position. The interesting part is that a “+/-“ can be calculated for each player based both within his assigned zone (IZ), and outside of that zone (OOZ). For example, if a player is +20 IZ, it indicates that based on the number of balls hit to his zone, he has made 20 more plays than the average defender at that position.

If the Cubs have simply been very fortunate defensively, we might expect to players very high IZ +/- (due to balls being hit right at people, or softly, candy hops, etc...), while subsequently seeing poor OOZ +/- (due to poor range, Ramirez’s lackadaisical attitude, etc…).

I have calculated this +/- data for the Cubs this year. All values are relative to the average position player between the 2004-2007 with a couple of changes:

  1. I used outfielder data only for 2006-2007, as the 2004-2005 numbers are skewed relative to these. I’m almost certain that changes to the OF zone rating system were instituted after 2005, but I will look into this further.
  2. I used only 2007 data for first baseman, as there seems to be a disconnect here also. Using the 2004-2007 data, all 1B in 2007 are significantly (> 10 plays OOZ) below average.

A couple of results are evident from the data:

  1. There does appear to be a significant distinction between the way the Cubs have handled their own zones vs. the areas just outside those zones. I conclude from this that the Cubs have indeed been lucky with regard to the outcome of balls in play. While the IZ, OOZ discrepancy may not be an ironclad indicator of BABIP luck, the key is that we see significant deviations from the average defense in both directions (+53.2, -38.6).
  2. The Mets are a useful comparison here in that, like the Cubs, they have a low BABIP. Unlike the Cubs, however, they have a number of players known for outstanding defense. The Mets are +71.8 IZ, +16.7 OOZ; they have played significantly better than average both in and out of their zones.
  3. If we simply combine the total “+/-“ for plays in and out of the Cubs’ zones, we find that the Cubs have made ~15 plays that the average team would not have. This amounts to about five points of BABIP, so obviously the translation is far from seamless, though I think useful nonetheless.

Lets hope the Cubs can leg out this pennant race, but lets just say I’m not as optimistic as when I began to look at the data.


Below is the data that Goldwater used:

No comments: