Catchin' Those Balls

by Andrew on Feb 15, 2007 9:11 PM PST in News Comment 4 comments

Defensive efficiency is a quality stat. A defense's job is to catch as many balls in play as possible, and that's exactly what defensive efficiency measures. For the most part, more balls caught equals better than. However, the stat does have some flaws. Some defenses face easier to field balls than others, and defensive efficiency, or DER, does nothing to account for that.

Since my mind tends to wander onto these subjects, and I had no real desire to study for my midterm, I came up with the creatively named stat "Advanced Defensive Efficiency" designed to fix some of the flaws present in defensive efficiency. The logic behind this stat is as follows:

Every type of batted ball in play has an average chance of being caught. For example, a line drive that stays in the park gets turned into an out 26.92 percent of the time. If you look at every ball in play that a defense faces, multiply it by the chance of it being turned into an out, then divide the sum of all those balls by the total number of balls in play, you get a number I call "Expected Defensive Efficiency" or eDER. This number represents the defensive efficiency a team would have if they caught each batted ball type an average amount of times. From here, advanced defensive efficiency is calculated by subtracting a team's DER by their eDER.

Before I present the numbers that I generated, I should warn you that this data is not entirely accurate. The exact amount of each batted ball type a team allows isn't available, just a percentage. Because of this, the final advanced defensive efficiency (aDER) numbers can be off by as much as half of a percent, and that's huge in a stat where the best and worst teams are separated by three percent. With that in mind, here are the aDER numbers for the 2006 season.

Rank	Old Rank	Team	DER	eDER	aDER
1	1	SDN	71.80%	70.16%	1.64%
2	2	DET	71.20%	69.70%	1.50%
3	6	CHN	70.50%	69.26%	1.24%
4	5	HOU	70.60%	69.49%	1.11%
5	9	SLN	70.40%	69.33%	1.07%
6	8	TOR	70.50%	69.67%	0.83%
7	13	SEA	69.80%	69.18%	0.62%
8	3	NYN	70.80%	70.30%	0.50%
9	18	ATL	69.20%	68.70%	0.50%
10	17	ARI	69.30%	68.81%	0.49%
11	4	NYA	70.60%	70.64%	-0.04%
12	14	MIL	69.40%	69.45%	-0.05%
13	21	CIN	69.10%	69.27%	-0.17%
14	10	ANA	70.30%	70.61%	-0.31%
15	25	CLE	68.60%	69.01%	-0.41%
16	7	SFN	70.50%	70.93%	-0.43%
17	26	BAL	68.50%	69.07%	-0.57%
18	22	COL	69.00%	69.60%	-0.60%
19	11	CHA	70.00%	70.67%	-0.67%
20	19	MIN	69.20%	69.98%	-0.78%
21	29	TBA	68.00%	69.21%	-1.21%
22	16	FLO	69.30%	70.55%	-1.25%
23	23	PHI	68.80%	70.12%	-1.32%
24	30	PIT	67.60%	68.95%	-1.35%
25	20	LAN	69.10%	70.51%	-1.41%
26	12	WAS	69.80%	71.24%	-1.44%
27	24	TEX	68.80%	70.26%	-1.46%
28	15	OAK	69.40%	71.11%	-1.71%
29	27	BOS	68.30%	70.33%	-2.03%
30	28	KCA	68.20%	70.36%	-2.16%

Abbreviations used:
Old Rank - The team's rank in defensive efficiency
DER - Defensive efficiency
eDER-Expected Defensive Efficiency
aDER- Advanced Defensive Efficiency

In 2006 The Dodger pitching staff allowed less line drives than any other pitching staff in baseball, which should have resulted in a high DER for the Dodgers. Instead, they finished 20th in baseball in DER which resulted in them tumbling further down the list in aDER. The big surprise here is Oakland, a team celebrated for its defense. They had the second highest eDER in baseball, yet finished only 15th in defensive efficiency, making them the third worst defense in baseball, and no where near the fourth to last. This conclusion seems to agree with the unremarkable rate2s Oakland put up at every position but third base and center field.

I like where this research is heading and I'd like to expand on it some more. There's two ways I can do this. First is getting actual numbers for batted balls instead of percentages. That will let me make actual conclusions from this data instead of saying "yeah, the Dodgers probably weren't very good defensively, but the data could also be flawed". The second is to introduce park factors into the equation. I think with Dodger Stadium's small power alleys line drives and fly balls get turned into outs more often, but I can't draw any real conclusions.

While individual defensive metrics are far from perfect, I think that team defensive metrics can be pretty well nailed down, and that aDER could be the best way to go about it.

Some other notes:

Percentage of time each type of batted ball in play is turned into an out:
Line Drives: 26.92%
Ground Balls: 74.8%
Outfield Fly Balls: 86.13%
Infield Fly Balls: 98.8%

Batted ball values from "What's A Batted Ball Worth?" by Dave Studenmund in the 2006 Hardball Times Annual.

0 recs | Comment 4 comments

Story-email Email Printer Print

Comments

batted ball type

While your post focuses on defense, the data about line drives, infield pop ups, etc. made me think of a similar issue with pitchers.

I'm wondering:

(1) has the availability of data on batted ball type settled the question of how true DIPS theory is? i.e. can we say, yes, K rate in the long run accounts for 95% of hit rate, or 80% or whatever?

(2) if there ARE significant repeatable skills in ability to manipulate hit rate independent of K rate, what are they? An ability to generate infield pop ups, for example? I've heard this is how Barry Zito allows consistently fewer hits than his K rate would suggest.

thanks.

by Alfredo Griffin on Feb 16, 2007 7:24 AM PST reply actions 0 recs

Re

I don't have the exact numbers with me, but I believe pitchers have been shown to have slight control over balls in play, but it still is mainly luck.

Infield pop ups are one of the more repeatable skills that a pitcher has.

by Andrew on Feb 16, 2007 7:57 AM PST up reply actions 0 recs

cumulative batted-ball data

I've got something for you I think. Using BPro's sortable statistics, you can add FB, LINERD, GB and POPUPS and calculate the percentages on your own. The raw data is there for 2005-2006.

I don't remember what James Click did for defensive efficiency offhand (it's been awhile since I read it and I'm at work at the moment) but you might want to check his work out in the BP archives.

Hopefully that helps; the BIS data and BP data are fairly similar, and BP has the raw totals.

by Marc Normandin on Feb 16, 2007 4:08 PM PST reply actions 0 recs

interesting

i found the post interesting. i wonder if this somehow relates to the discussion of Ethier's second half decline. I read somewhere that he had an unusual amount of line drives that fell for singles, initially, and then they started to getting caught. I don't know.

The point is, I have this feeling that classification of the types of ball in play is where the real important work is at for this kind of analysis. (Even defining the categories themselves is critical). I am sure you are aware of this.

the funniest part of your post, as a dodger fan, is your reference to the angels as 'ANA' and not 'LAA'. why not, may i suggest, 'OCA'? ;-)

by cb504 on Feb 19, 2007 6:30 AM PST reply actions 0 recs