I noticed some discussions about the clutch rating earlier. I have a concern about it.
I am concerned that the clutch rating may overcompensate for differences in ERA for relievers.

First, as I understand it, the clutch rating is designed to account for differences in ERA between pitchers who are otherwise equal. So, if two pitchers both give up the same numbers of walks and hits and get the same numbers of strikeouts per inning, but one has a substantially higher ERA than another, the clutch rating system assumes that he gave up more “big hits” (i.e. with runners on base). Thus, in game play, the pitcher with the lower clutch rating is penalized with runners on base (every time there are runners on base).

This system seems well-designed for starters who pitch 175-225 innings. I think it may overcompensate with relievers, however. Let’s take the example of three very good relievers from 2002:

NAME G IP H R ER BB SO W L Hld Sv ERA Rating Clutch
J. Romero, Min 81 81.0 62 17 17 36 76 9 2 33 1 1.89 8.4 8.1
J. Smoltz, Atl 75 80.1 59 30 29 24 85 3 2 0 55 3.25 8.6 4.6
L. Hawkins, Min 65 80.1 63 23 19 15 63 6 0 13 0 2.13 9 7.6

As you can see, these three pitchers, in virtually the same number of innings, compiled VERY similar stats in terms of pitching success. They gave up almost identical numbers of hits, and had comparable numbers of walks vs. strikeouts. This is reflected in their very similar (and high) PB ratings. Smoltz, however, had an ERA nearly a run and a quarter higher than Hawkins and nearly a run and a half higher than Romero. As a result, his clutch rating is a sub-par 4.6, where Hawkins’ and Romero’s are both very good.

Now let’s consider the reason for this. On April 6th, against the Mets, Smoltz pitched 2/3 of an inning, allowing 6 hits, 2 walks and giving up EIGHT earned runs while striking out 2. If you back this single game out of Smoltz’ line for the season, the numbers for the three pitchers look like this:

NAME G IP H R ER BB SO W L Hld Sv ERA Rating Clutch
J. Romero, Min 81 81.0 62 17 17 36 76 9 2 33 1 1.89 8.4 8.1
J. Smoltz, Atl 74 79.2 57 22 21 22 83 3 1 0 55 2.37 ?? ??
L. Hawkins, Min 65 80.1 63 23 19 15 63 6 0 13 0 2.13 9 7.6

At that point, Smoltz’ rating probably goes north of 9, and I would think his clutch rating would have to go to at least 7. Does it really make sense to penalize Smoltz that significantly EVERY time he pitches with runners on for the difference that was caused by one bad outing?

In short, I think that relievers’ ERAs are flaky. They can vary so much based on 1 or 2 bad outings that I think it is unfair to base game play differentials on them.