I love lounging around on summer days and watching baseball, but I also love to read great baseball work as well as do my own analysis and writing. In fact, I’d say I spend a solid 72% of my days thinking about baseball (don’t ask about the other 28%). My obsession has resulted in ranking 39th lifetime in the National Fantasy Baseball Championship and what seems like a lifelong dream of writing for BP.

The Gibbons Conundrum: Effects of Defense on Pitching

Jeremy Guthrie, rookie pitcher for Baltimore, out of nowhere, had been taking the league by storm when he stepped onto the hill on July 12th, 2007 against the White Sox. He easily recorded the first couple outs before Jim Thome hit a drive to deep left. The ball barely cleared the wall and most defenders would have caught it, but Jay Gibbons couldn’t haul it in and a homer as well as an extension of the inning was placed on Guthrie’s line. The Sox kept hitting before a hit straight at Gibbons went through his legs, preventing him from throwing out the runner at the plate. The next batter to hit a can of corn, as Hawk Harrelson might say, was Rob Mackowiak, again dropping in an easy flyball near Gibbons that most fielders would have gotten to. When the inning was over, four runs had crossed the plate on Guthrie’s watch, three of them earned but all of them were due to the negligence of Gibbons, prompting Guthrie fantasy owners everywhere to post messages on their league board, like, “Hey, Jay Gibbons, @%*# YOU!”

It is no secret among the initiated that errors are a poor way to judge whether defenders are effective. In recent years, more effective ways to judge defense have been developed-some such as John Dewan’s Plus/Minus and Michael Lichtman’s UZR use detailed play-by-play information. Major league clubs have also worked on their own interior evaluation systems because they realize that preventing runs via defense can be nearly as important as pitching and hitting. Dewan made this point in a recent Sports Illustrated article: “The difference between the best defensive team in baseball and the worst defensive team in baseball is about 130 runs. On the batting side, the difference between the best and the worst team is about 260 runs. To think that the value of fielding is worth as much as half the value of offense, I don’t think anyone would have thought that. That’s a significant number.”

What Dewan doesn’t note is that it is his number. His defensive system is well respected and certainly better than the archaic scorekeeper stats, but how do we know whether he’s evaluating defense correctly? And it leaves us with questions about how many runs extra or less a specific pitcher is giving up due to his defense. We know defense is important, but exactly how important is an intriguing area for study.

The easiest and most accurate way to judge team defense has long been through defensive efficiency (DEF_EFF or DER)-a measure of how many balls put into play were converted into outs by a team’s defense. Unfortunately, this metric doesn’t help us evaluate individual players, but since pitchers have little control over the destiny of balls in play, it is a pretty good measure of team defense.

It also has quite an effect on both the RA and ERA of players and teams. In order to study the effect, I took each pitcher’s performance in the MLB over the last 5 years and looked at the Deff_Eff, ERA, and RA of each one (weighted by the number of innings each pitcher threw). I then separated these pitchers into two groups, ones who had good defensive support (better than the .6924 average) and ones who happened to get the more inept (or hungover) group of fielders. Some profound results came up. The better group had a Deff_Eff of .718 with a 3.85 ERA and 4.16 RA. The less fortunate group had a Def_Eff of .666 with a 4.97 ERA and 5.41 RA. This works out to each .001 or thousandth of a point of Deff_Eff being worth .0215 runs of ERA and .024 runs of RA. That may not sound like much, but when you consider that the best and worst teams are often 40 thousandths of a point of Deff_Eff, it is quite a bit. You’ll also note that the errors affecting ERA really don’t make much difference.

What does all this mean? For one thing, it means that Tampa Bay team last year had a much better ERA than the year before because they had a great defense posting a .710 Deff_Eff, not because their pitchers (much the same as the year before) learned how to pitch. It also means that Texas pitching would be somewhat less horrid if their fielders got to balls a little more often than 67% of the time.

Let’s take a look at how team ERA would have fared in 2008 if they had average pitchers, but were adjusted only by their Deff_Eff:

TEAM  DEF_EFF    ERA  Adjusted ERA
TBA     0.71    3.82   4.04
CHN     0.705   3.87   4.15
TOR     0.704   3.49   4.17
OAK     0.7     4.01   4.25
BOS     0.699   4.01   4.27
NYN     0.698   4.07   4.30
MIL     0.698   3.85   4.30
HOU     0.698   4.36   4.30
SDN     0.696   4.41   4.34
PHI     0.696   3.88   4.34
SLN     0.695   4.19   4.36
ATL     0.694   4.46   4.38
FLO     0.693   4.43   4.40
ANA     0.692   3.99   4.42
LAN     0.691   3.68   4.45
KCA     0.69    4.48   4.47
WAS     0.689   4.66   4.49
BAL     0.688   5.13   4.51
MIN     0.687   4.16   4.53
CHA     0.686   4.06   4.55
ARI     0.686   3.98   4.55
CLE     0.686   4.45   4.55
SFN     0.685   4.38   4.58
DET     0.685   4.9    4.58
NYA     0.682   4.28   4.64
SEA     0.682   4.73   4.64
COL     0.678   4.77   4.73
PIT     0.675   5.08   4.79
CIN     0.673   4.55   4.83
TEX     0.67    5.37   4.90

Clearly, the teams with better Deff_Eff had better ERA‘s but we can also see that their ERA‘s adjusted only by defense weren’t far off from their actual effectiveness. The caveat here is that the ballparks have some effect, but most of what you see here is the defense in action. In terms of pitching, the same pitcher could end up with an ERA higher or lower by almost a whole run depending on the team for which he pitches. If you spent last year wondering why the Yankees or Tigers started with promising pitching staffs that finished poorly, you can see the holes in which their fielders were putting them.

Let’s go back to Dewan’s quote. He posited that the difference between the best and worst fielding teams is 130 runs. In the case of DER, it should have been .86 runs per 9 innings, or 138 runs over a full season (roughly 1,450 innings). This helps to support not only the validity of Dewan’s work but also the whole idea that defense weighs heavily on pitching stats. Next time you like a pitcher for the upcoming season, do what I do and consider whether he has a team of Jay Gibbon’s bad days behind him.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
The guy who created UZR is actually named Mitchel Lichtman, not Michael Lichtman. The lack of fact-checking and editing is the most immediately noticeable difference between the BP Idol submissions and the work by the pros. I suppose that can also be read as a compliment to the authors, since the work is pretty good, just not polished.
Don't forget, the BP writers also have a pretty darn good editor. All I had for my submission was my pet cat.
I hear ya. My cat's always getting "our" and "hour" confused.

I'm starting to think she's not very bright. :-/
I'm confused. The higher defensive efficiency scores at the top of the chart lead to the Adjusted ERAs being higher than the straight ERAs. But at the bottom of the chart, Cincinnati's ERA is also adjusted upward, even with a very poor defense. What am I missing?
I didn't adjust the actual ERA at all. I was just putting it there to show what it was. The adjusted ERA for each team takes the league average ERA and alters it according to the team's defense. The idea is that if each team had an average pitching staff in every way, we get to see how the team's ERA would be altered by the defense alone.
1. How do you adjust the team's ERA based on DEF? Are you just calculating the change in ERA based on the number of baserunners a different DEF would cause? That needs an explanation.

2. I was also wondering about the validity of DEF being all due to fielding.

3. I wanted to know precisely how much Tampa Bay's DEF and ERA improved over the year before.

4. The chart could have been nicer looking - trying to read the All Caps abbreviations for the teams was not enjoyable.

5. Otherwise, nice work, thanks. Very good intro.

Prior to the chart I found that each thousandth of a point of DER is worth .0215 runs of ERA. Taking that into account, it is easy to make the chart by finding the difference between team DER and the average DER and then adding the corresponding ERA to the average ERA. Hope that helps.
I still don't understand how the Reds' adjusted ERA went up, when they had one of the worst DEFs. Is it just a typo?
The adjusted ERA takes the league average ERA and applies defense. The Reds start with an average ERA (like every team), which goes down to below average when you apply the defensive debit. Their pitching staff was also pretty bad which is what makes their real-life ERA worse than the defense-only adjusted ERA.
But it's NOT worse, according to the chart. That's what has me confused.
Very good article, powerful chart. Never really considered that if you put the Rangers pitchers in front of the Rays defense it'd be worth 1.55 runs of ERA.
But that would be assigning 100% of the DER to the defense. Voros McCracken concluded this, but further research by others suggests it's in the range of 20% pitchers 80% fielders. Maybe it wasn't the writers intention, but I was looking for a dtermination of this pitcher/fielder split and didn't find it. And ballparks do also have an influence.
When you say "average" pitchers, do you mean every stat is adjusted, of just the defense independent ones, like K-rate, G/F, etc.?
I'm calculating it by starting with a league average ERA for the pitching staff--then adjusting according to defense.
Good article except "The caveat here is that the ballparks have some effect, but most of what you see here is the defense in action" is just begging for more analysis. Or else, you get comments similar to donwinningham's where the implication is made that the Rangers pitchers would have an ERA exactly 1.55 runs better if it had Tampa Bay's defense. I'm not picking on donwinningham, but illustrating the point that if this article is an evaluation of Dewan's sytem, then how much ballparks affect defense has to be taken into consideration (though I recognize there was also a word limit on submissions).
I definitely want to do another article on park effects. For most teams it doesn't make much difference but it has huge effects on Boston and Oakland, for example. However, the outer limits of difference in ERA will stay relatively the same.
I'm adding my judging comment to each article:

Oakchunas, Brian -- 7. This is a solid research piece with some decent enough writing. He goes right after it, doesn't get cute, but relies a bit too much on the central chart to carry his argument. I'd have liked to see a bit more explication here, but his process is solid and the topic is a bold choice.
I like this topic, but I'm not sure the method gets the right conclusion. My initial guess is that it's biased towards finding a larger effect for defense.

Say that there are teams that are primarily focused on run prevention. These teams are going to, on average, get better pitchers and better defenders. So teams with good defenses are going to look good for 2 reasons: 1) defense saves run; 2) they have better pitchers. A simpler way to say this is that you found the correlation, but it doesn't imply causation. (To ramble a bit - We care about what should happen to run prevention when a team improves its defense. You found the run prevention of teams with different quality defenses.)

I can think of other stories too (some even go in the other direction), but the larger point (correlation vs. causation) still holds.

Here's an idea that I think solves this problem and gives you a fantastic topic for the future: Do this by pitcher. Look at the relationship between a pitcher's ERA and the DER he got in that year. Look at how the pitcher's ERA changes when the DER changes year-to-year. That would be pretty cool. This keeps the quality of the pitchers "fixed."

Needless to say, the fact that I even spent time thinking about this means this was a really interesting article.
I liked the idea, but didn't realize the method behind the calculations until it was made clear in the comments section. I guess I always found it hard to describe mathematical calculations using only words. Perhaps a more detailed explanation using the formulae, and words to explain them, would have been better.
I also didn't understand the chart the first time through, but seeing it now, this article is very informative.

While reading the BP annual, I often get hung up that BP assumes it's defensive stat is accurate when I myself have no such confidence in it. Maybe we actually are at the point where we can rely on these new defensive stats.
I ignore some of BP's defensive stats to be honest. I got tired of seeing a fielder rated a 10 (for 10 runs above average) for defense, but being called an inferior fielder in the comment section. It's a similar pet peeve I have to some fantasy baseball magazines where it's completely obvious that whoever wrote the comment for a player didn't look at the player's projection for the upcoming year.