Stephen discussed starting pitchers today. I thought I would take a quick look at how Brad Mills is using his bullpen. Fangraphs used a quick and dirty approach to looking at relief pitcher usage by comparing the game leverage index for teams' top three relievers to CHONE ERA. This may or may not be a good methodology; but it's intended to gauge the extent that teams are putting their best relief pitchers in high leverage situations. The Astros fell in the broad middle group of major leagues teams, according to this measure.
I thought it would be interesting to look at how Astros relief pitchers' performance so far stacks up against their game leverage index. (Game leverage index is the leverage index when the pitcher enters the game.) For measures of current performance I used tERA from Fangraphs, which is a more advanced fielding neutral pitching measure that recognizes the type of batted balls and probability of various outcomes. I also used a combination of FIP and regular ERA as an alternative measure. In theory, tERA should be the best measure of peformance; I have qualms about using either ERA or FIP to measure relief pitcher performance, and I split the difference by using a calculated value which is the average of FIP and ERA. Since Gervacio was on the DL for much of the season so far, I haven't included him in the table. But, when he was available, Mills apparently liked using him, since he had the fourth highest gLI in the bullpen. Gervacio's tERA is a sparkling 1.84 and his average FIP-ERA is 3.42.
We'll get to the results after the jump.
Before I lay out the table, I'll give you my basic conclusion, which is that General Mills has been pretty good at using his best performing relievers in the highest leverage situations. Or, alternatively, you can give the pitchers credit, and say that the relief pitchers have been good when they have been placed in higher leverage positions. I doubt that this is all that surprising. After all, relief pitchers were assigned roles based on expectations regarding their talent. But it's reassuring when it turns out that high leverage relievers pitch to their expectations.Let's try nit picking things, though. Arguably, Lyon, who has been used in the second highest leverage situations, hasn't been the second best reliever so far. Part of this observation is due to the fact that Lyon had some rough outings initially, but has been very tough as the set up man in recent outings. Sampson has performed at a high level so far, and could be viewed as the best relief pitcher not named Lindstrom. Mills has used him in the third highest leverage situations so far. If he were given the set up role, I suspect Sampson would do a good job. On the other hand, I think that the Astros are fortunate to have a strong reliever like Sampson who can give Lyon a breather in the set up role, if necessary. In addition, Sampson is the best pitcher to perform the "fireman" role, when help is needed in a critical late game situation with runners on base. Sampson's pitching repertoire is well suited to inducing the double play. So, if it's the 7th inning, with runners on the corners and 1 out, Sampson may be the best choice for getting the optimal outcome. If it's the same inning, with runners on 2d and 3d with no outs, I might favor Gervacio, who can get the strike out, depending on the game score and the batter.
As we have noted in other articles, Mills seems to use Byrdak more as a normal reliever than as a LOOGY (Lefthanded One Out Guy). Overall, Byrdak has functioned fairly well when he is used as a normal middle reliever. His control and nibbling are still problems,which could prove to be his Waterloo in the future. So far, Byrdak is producing as many walks as strike outs, which is a bad sign. However, it is early and the sample size is tiny--so I'm not reaching any conclusions yet. Brian Moehler is an odd case. He is a former starter who has posted good pitching numbers out of the pen so far--at least based on my measures in the table. Moehler's FIP is better than his ERA, which is unusual for him. However, he has pitched in the lowest leverage situations. I think this is a case where Mills is still trying to figure out how Moehler fits into the bullpen. He has been used in the late innings on occasion so far, but I think he probably is kept in reserve for long relief outings or multi-inning outings in the 5th or 6th inning; and because the starters have pitched well, he hasn't received many opportunities. The below shows the two low leverage relievers.
| Low leverage | tERA | Avg. FIP+ERA | gLI |
| Fulchino | 6.46 | 6.46 | .39 |
| Moehler | 3.33 | 3.34 | .22 |
The table below shows the higher leveraged relievers in descending order of leverage.
| high leverage | tERA | AvgFIP+ERA | gLI |
| Lindstrom | 2.77 | 3.42 | 1.65 |
| Lyon | 3.35 | 4.54 | 1.14 |
| Sampson | 3.09 | 2.00 | 0.98 |
| Byrdak | 4.44 | 4.47 | 0.8 |
| Lopez | 4.13 | 4.28 | 0.7 |