Tuesday, June 01, 2004


"Compare [D-Backs bullpen use] to other ballclubs and other bullpens and other managers and you'll find we are not out of line. We do not overuse our guys." -- Bob Brenly

As has been mentioned here and elsewhere, the drumbeats regarding Brenly's use of the bullpen are getting louder -- see this article for Brenly and Joe Jr.'s responses.

When complaints reach fever pitch, that's when I find it useful to see if the complaints hold any water. So, I'm going to use Baseball Prospectus' Relievers' Run Expectation report to look at NL bullpens.

Let's start off by looking at their rankings of teams by "Adjusted Runs Prevented" (ARP), a stat designed to evaluate "the reliever handles his inherited runners, and the support the reliever gets when he turns runners over to others."

Team ARP
PIT 22.2
LAD 20.8
PHI 19.9
ATL 18.4
SDP 17.5
HOU 13.0
FLA 6.1
MON 5.5
COL 3.2
NYM 1.8
CHC 0.5
SFG -2.4
CIN -6.0
STL -6.8
MIL -7.1
ARI -7.6

Yep, our staff has not performed well. (In a related development, we understand that the sun rose in the east this morning.)

But is this poor performance due to overuse? Let's take a look at staffs ranked by number of relief innings.

Team ARP Relief Innings
STL -6.8 134.3
MON 5.5 134.7
CHC 0.5 140.3
ATL 18.4 142.3
FLA 6.1 144.7
PHI 19.9 152.0
SDP 17.5 153.7
CIN -6.0 155.7
HOU 13.0 156.7
PIT 22.2 157.7
LAD 20.8 158.0
SFG -2.4 161.7
NYM 1.8 166.3
COL 3.2 167.7
ARI -7.6 171.0
MIL -7.1 179.3

Hey, Arizona's no longer in last place! And their placement their with Milwaukee doesn't make them look good. But look at St. Louis -- they never use their bullpen. (Apparently because their bullpen is almost as lousy as ours.) It's been too long since I've taken all my statistics classes, so I'm not going to bother running a regression here, but if there's a relationship between bullpen use and bullpen effectiveness, I'm not seeing it.

How about the fact that not everybody's played the same amount of games... This chart is ranked by innings per game played.

Team ARP Innings Inn./Game
STL -6.8 134.3 2.69
MON 5.5 134.7 2.69
CHC 0.5 140.3 2.81
FLA 6.1 144.7 2.84
ATL 18.4 142.3 2.85
SDP 17.5 153.7 3.01
CIN -6.0 155.7 3.05
PHI 19.9 152.0 3.10
HOU 13.0 156.7 3.13
LAD 20.8 158.0 3.22
SFG -2.4 161.7 3.23
NYM 1.8 166.3 3.33
ARI -7.6 171.0 3.35
COL 3.2 167.7 3.35
PIT 22.2 157.7 3.36
MIL -7.1 179.3 3.66

Arizona climbs to #13 by about 1/1000th of an inning above Colorado. Again, no visibly apparent statistical correlation.

Finally, perhaps the problem is that individual relievers are being overtaxed. There are two ways to do this: 1) look at pitchers over a certain inning count and compare their effectiveness over time, adjusted for park effects, etc., or 2) divide the number in the last column by the number of relievers on staff, as counted by Baseball Prospectus for the entire season.

I will choose #2, of course. The table below shows teams ranked by the average innings pitched per game by the average reliever.

Team ARP Innings Inn./Game Inn./Game/Reliever
SFG -2.4 161.7 3.23 .248
CHC 0.5 140.3 2.81 .255
STL -6.8 134.3 2.69 .269
SDP 17.5 153.7 3.01 .274
MIL -7.1 179.3 3.66 .282
ARI -7.6 171.0 3.35 .305
PHI 19.9 152.0 3.10 .310
FLA 6.1 144.7 2.84 .316
COL 3.2 167.7 3.35 .335
MON 5.5 134.7 2.69 .336
PIT 22.2 157.7 3.36 .336
CIN -6.0 155.7 3.05 .339
HOU 13.0 156.7 3.13 .348
ATL 18.4 142.3 2.85 .356
NYM 1.8 166.3 3.33 .370
LAD 20.8 158.0 3.22 .403

What does this tell me? Again, not much. Milwaukee and San Francisco leap up because they've used 13 different pitchers in relief situations this year. Arizona's used 11 different pitchers and ranks 6th on this table. Atlanta and the Dodgers each have used just 8 pitchers, the Mets 9.

There are many problems with the methodology here (e.g., sample size, definition of relief pitcher, inability to look at results over time as "overuse" increases), but I will use the data anyway to draw a conclusion:

The D-Backs bullpen isn't very good because the bullpen isn't very good.

Meaning, Mantei and Valverde and Choate haven't pitched well not because they've been overused but because they haven't played well. Even if the bullpen is "overused," the data don't seem to point to overuse having any statistical effect.

I'm not saying Brenly makes odd bullpen decisions -- maybe he does point to the bullpen in left field earlier than he should -- but based on this, I'm not convinced overuse is the reason for their poor performance thus far.

The D-Backs bullpen isn't very good because the bullpen isn't very good.Doesn't get any simpler than this. Nice piece, Stefan.

Like most, I thought their bullpen would be a huge asset this season and based many of my expectations on that assumption. I fully expect positive regression to kick in eventually, but it won't make much of a difference at this point.
Thanks for the comment, Robert.

You and I do not always agree on how much blame management should get for their player decisions (I do tend to not give them as much blame as you), but I will say that management has done a poorer job with evaluating pitchers than with the rest of the squad. From what I can see in Tucson, et al., I think our staff of the future will have to come via free agency and trade.

And I know what you think about that possibility. ;-)
Post a Comment