Monday, December 10, 2007

Team Run Prevention in 2007

I could probably talk about the big trade for the rest of the winter (and I will) but today I want to get back to the sabermetric review of the 2007 season. Evaluation of the offensive part of the game is fairly straight forward but the defensive side of the game is more complex. The goal of every baseball team when they are not at bat is to give up as few runs as possible. This, of course, is done with pitching and fielding. The problem is how to separate pitching from fielding so we can evaluate each one of them in isolation. Because pitching and fielding are so intertwined this is not an easy task. However, sabermetric advancements are being made.

The first person known to quantify team fielding ability in a useful way was Bill James in his abstracts published in the 1980s. He developed a measure which he called Defensive Efficiency Ratio (DER). DER is the percentage of times batted balls are turned into outs by the team's fielders, not including home runs. There are different versions of the formula but the one now most commonly used is DER=(BFP-H-K-BB-HBP-0.6*E)/(BFP-HR-K-BB-HBP) where BFP = batters faced pitcher, H=hits allowed, K=strikeouts, BB=walks allowed, HBP=hits batsmen and E=errors.

For years, the only attempt to isolate pitching from fielding was ERA which was supposed to remove the effect of fielding errors from runs allowed. ERA is flawed, in part, because it is dependent on the whims of different official scorers. More importantly, it does not take into account the ability of fielders to get to balls.

In 2000, Voros McCracken published the results of his study which showed that there is little difference among major league pitchers in their ability to prevent hits on balls hit in the field of play. This conclusion was met with much skepticism even in the sabermetric community. His theory has been tested rigorously and weakened somewhat but his main concept has so far been shown to be true. That is, hits allowed by pitchers are more a function of team fielding ability than the ability of the pitcher.

The idea is that the contribution of pitching to overall team defense can be derived, for the most part, from walks allowed, strikeouts, homeruns allowed and hits batsmen issued. McCracken referred to those as Defense Independent Statistics (or DIPS). Tom Tango derived a statistic from DIPS called Fielding Independent Pitching (FIP) ERA. FIP ERA measures pitcher performance essentially independent from fielding. The formula is HR*13+(BB+HBP)*3-K*2)/IP plus a league specific factor to make it equivalent to ERA.

Looking at team statistics over the past 5 years, it is seen that 92% of the variance in team runs allowed per game (RA/G) is explained by FIP ERA and DER. This is a very high correlation and it supports the theory that FIP ERA and DER explain most of run prevention. Furthermore, FIP ERA, by itself, explains 64% of the variation in RA/G and that DER, by itself, explains 48% of the variation in RA/G. This tells us that pitching is more important to run prevention than fielding but that fielding is also essential. In short, what Bill James has always said is true: “Much of what we think of as pitching is actually fielding”.

So, we have DER to measure overall team fielding and FIP to measure overall team pitching. Table 1 below shows the runs allowed (RA), ERA, FIP and DER for each American League team in 2007. It also shows how the teams rank on those statistics. The raw data were extracted from The Hardball Times site.

It can be seen from the table that the Tigers were 9th in the league in run prevention in 2007 allowing 4.92 runs per game. game. Their pitching was not good as they finished 11th in the league in with a 4.72 FIP ERA. in 2006, they finished 3rd at 4.36. Their pitching was saved somewhat by their fielding which finished 4th with a .693 ERA in 2007. However, their glove work was not as stellar as it was in 2006 when they finished first with a .704 DER.

I will note that DER is not purely a fielding stat because pitchers do have some control over batted ball types (line drives, ground balls, etc) and ballparks also play a role. However, it is a useful fielding measure and it’s also simple and accessible. In future articles, I will touch upon more sophisticated fielding stats.

Table 1: AL Team Run Prevention in 2007


Team

RA/G

RA/G Rank

ERA

ERA Rank

FIP

FIP ERA Rank

DER

DER Rank

BOS

4.06

1

3.87

1

4.24

3

.706

1

TOR

4.31

2

4.00

2

4.35

5

.706

2

CLE

4.35

3

4.05

3

4.12

1

.687

8

MIN

4.48

4

4.15

4

4.45

6

.689

6

LAA

4.51

5

4.23

5

4.20

2

.681

12

OAK

4.68

6

4.28

6

4.25

4

.694

3

NYA

4.80

7

4.49

8

4.59

9

.690

5

KC

4.80

8

4.48

7

4.58

8

.685

9

DET

4.92

9

4.57

9

4.73

11

.693

4

SEA

5.02

10

4.73

10

4.48

7

.675

13

CHA

5.18

11

4.77

12

4.60

10

.684

10

TEX

5.21

12

4.75

11

4.89

14

.683

11

BAL

5.36

13

5.17

13

4.85

13

.688

7

TB

5.83

14

5.53

14

4.78

12

.657

14

Ave

4.82


4.51


4.51


.687


1 comment:

  1. This comment has been removed by a blog administrator.

    ReplyDelete

Twitter

Blog Archive

Subscribe

My Sabermetrics Book

My Sabermetrics Book
One of Baseball America's top ten books of 2010

Other Sabermetrics Books

Stat Counter