ERA has traditionally been used to evaluate overall pitching performance but it is problematic because it does not separate pitching from fielding. More and more, I've been using FIP ERA (or Fielding Independent ERA) which is calculated using only statistics over which a pitcher has control - strikeouts, bases on balls, hits batsmen and home runs. FIP is more predictive or more stable from year to year than ERA.
In last year's article on repeatable pitching skills, I showed that SO/IP (correlation=.77) and BB/IP (correlation=.67) are very repeatable from one year to the next. Home runs per IP is less consistent (.38) and it is also park dependent. For this reason, many analysts are now using ground ball percentage (percent of batted balls which are hit on the ground) more than home runs. Ground ball percentage is relatively consistent (.73) and is also not very dependent on ballpark. One disadvantage of ground ball rate is that it is not an actual outcome. Getting a batter to hit a ground ball is fielding independent but getting the ground ball to turn into an out is not.
The statistic QERA, developed by Nate Silver at Baseball Prospectus, uses GB% instead of home runs:
QERA= (2.69 - SO% x .34 + BB% x 3.88 - GB% x .66)^2
While QERA is more predictive of future performance than ERA (.51 versus .33), I have not found it to be more predictive than FIP (.54). Still, GB% by itself is a useful tool which helps to describe an important pitching skill.
While a ground ball does not always produce a good result for a pitcher, it’s much less likely to produce a bad result than a line drive or a fly ball. Table 1 (data taken from a discussion at Fangraphs) shows that ground balls result in a much lower slugging average than either line drives or fly balls.
Table 1: Slugging Percentage By Batted Ball Type
Event | Slg |
Line drive | .978 |
Ground ball | .220 |
Fly ball | .494 |
Futhermore, Table 2 (taken from the The Hardball Times Annual 2007) indicates that a ground ball is less likely to contribute to runs than either a fly ball or a line drive. The exception would be an infield fly but those are much less common than outfield flies. Whatever way you look at it, it is clear that the ability to get batters to hit the ball on the ground is a good thing.
Table 2: Run Impact of Batted Ball Type
Event | Run Impact |
Line Drive | .391 |
HBP | .355 |
Walk | .355 |
Outfield Fly | .192 |
Intentional Walk | .075 |
Ground ball | .045 |
Bunt | .021 |
Infield fly | -.088 |
Strike out | -.113 |
Table 3 shows how Tigers starting pitchers (plus Dontrelle Willis) ranked in 2007 in ground ball percentage. Table 4 lists the top 20 ground ball pitchers in the American League in 2007. After having three starters (Kenny Rogers, Jeremy Bonderman and Nate Robertson in the top 12 in 2006, the Tigers top ground ball pitcher in 2007 was Bonderman who finished 16th at 48%. Rogers also had a 48% ground ball rate but was limited to 63 innings due to injuries. Nate Robertson had a league average ground ball rate (45%) while Justin Verlander finished at 41%.
The raw data for Tables 3 and 4 were abstracted from The Hardball Times database.
Table 3: Ground ball Rates for Tigers (plus Dontrelle Willis) Starters in 2007
Name | IP | GB% | ERA | FIP | |
16 | Bonderman | 174.3 | 0.48 | 5.01 | 4.22 |
23 | Willis | 205.3 | 0.46 | 5.17 | 5.10 |
33 | Robertson | 177.7 | 0.45 | 4.76 | 4.73 |
36 | Durbin | 127.7 | 0.44 | 4.72 | 5.73 |
45 | Verlander | 201.7 | 0.41 | 3.66 | 4.09 |
. | Jurrjens | 30.7 | 0.38 | 4.70 | 5.40 |
. | Maroth | 78.3 | 0.43 | 5.06 | 6.53 |
. | Miller | 64.0 | 0.49 | 5.63 | 5.41 |
. | Rogers | 63.0 | 0.48 | 4.43 | 5.13 |
GB% Rank | Name | Team | IP | GB% | ERA | FIP |
1 | Carmona | CLE | 215.0 | 0.64 | 3.06 | 4.05 |
2 | Hernandez | SEA | 190.3 | 0.61 | 3.92 | 3.83 |
3 | Wang | NYA | 199.3 | 0.58 | 3.70 | 3.92 |
4 | DiNardo | OAK | 131.3 | 0.56 | 4.11 | 4.93 |
5 | Loe | TEX | 136.0 | 0.56 | 5.36 | 4.67 |
6 | Burnett | TOR | 165.7 | 0.55 | 3.75 | 4.44 |
7 | Tavarez | BOS | 134.7 | 0.54 | 5.15 | 4.79 |
8 | Westbrook | CLE | 152.0 | 0.54 | 4.32 | 4.37 |
9 | Halladay | TOR | 225.3 | 0.53 | 3.71 | 3.65 |
10 | McGowan | TOR | 169.7 | 0.53 | 4.08 | 3.82 |
11 | Gaudin | OAK | 199.3 | 0.51 | 4.42 | 4.71 |
12 | Cabrera | BAL | 204.3 | 0.50 | 5.55 | 5.06 |
13 | Ramirez | SEA | 98.0 | 0.48 | 7.16 | 5.60 |
14 | Litsch | TOR | 111.0 | 0.48 | 3.81 | 5.23 |
15 | Bedard | BAL | 182.0 | 0.48 | 3.16 | 3.33 |
16 | Bonderman | DET | 174.3 | 0.48 | 5.01 | 4.22 |
17 | Pettitte | NYA | 215.3 | 0.48 | 4.05 | 4.00 |
18 | Silva | MIN | 202.0 | 0.48 | 4.19 | 4.35 |
19 | Beckett | BOS | 200.7 | 0.47 | 3.27 | 3.22 |
20 | Blanton | OAK | 230.0 | 0.47 | 3.95 | 3.59 |
No comments:
Post a Comment