Starters v. Relievers, Part II

The other day we were looking at how the Red Sox starters and relievers compare. The results were rather disconcerting: Rudy Seanez has better numbers than the starters?
We watch the games, we know he sucks. Something must be wrong. A few readers wrote in with some ideas. We'll start with inherited runners. Basically when a reliever lets an inherited runner score, that run is charged to whoever let the runner get on base and the reliever's ERA goes unscathed. Some readers saw this as giving Rudy Seanez a free pass for sucking and allowing tons of inherited runners to score.

We could just look at how many inherited runners a reliever strands vs how many inherited runners he has to deal with. A problem with that approach is that it doesn't make any distinction between entering the game with a runner on first and 2 outs and entering the game with a runner on third and no outs. Obviously we want to give a reliever more credit for succeeding in the later situation.

Here are the Red Sox relievers sorted by inherited runs prevented:

NAME                  IP       IRP
Craig Hansen          28.7     1.8
Mike Timlin           46       1.4
Keith Foulke          32       0.9
Jonathan Papelbon     61       0.8
Jermaine Van Buren    11.7     0.5
Kyle Snyder           28.3    -0.1
Lenny Dinardo         25.3    -1.3
Javier Lopez           8.7    -1.9
David Riske            9.7    -2.5
Manny Delcarmen       37.3    -4.1
Rudy Seanez           45.3    -6.6
Julian Tavarez        59.7    -6.9

IRP is a stat that looks at the situation when a reliever enters the game and how many runs a league average reliever would allow before getting out of the inning and subtracts the number of runs the reliever actually gave up. What's left are runs that would have scored on average but didn't because of the pitcher's work. Negative numbers are the number of runs that reliever let score that wouldn't have scored with the league average reliever in there. (#s are for 2006).