As you know, several teams have already made a strategic change of closers, due to injury or ineffectiveness. But can we tell if the number of changes by this point in the season is actually unusual?
I looked at the first 30 games for each season since the last expansion, and counted the number of pitchers with a “closer save” (1 IP or less) in that span. The following table shows the number of such closers with at least one such save, the number with two or more, and the average saves per closer.*
| Year | 1+ | 2+ | Sv. /Cl.* |
| Avg 1998-2011 | 47.6 | 32.7 | 4.0 |
| 2012 | 55 | 35 | 3.7 |
* Saves per closer is meant to curb false impressions from the year-to-year variance in total closer saves. The more saves there are, the greater the chance that a team will use a secondary closer simply because the lead man is overworked. If the total number of closers spikes in one year, but the average shows less change, that would tend to indicate more secondary closers rather than more strategic reassignments. At least, that’s my theory; it’s imperfect, but it’s the best I could think of.
From that snapshot, it looks like this year might be unusual. The number with 1+ is 16% above average, the number with 2+ is 7% above average; and the average per closer is 8% below average, suggesting that the rise in closers isn’t merely due to more available saves.
But what if we focus only on the past four years?
| Year | 1+ | 2+ | Sv./Cl. |
| Avg 2008-2011 | 50.8 | 35.5 | 4.0 |
| 2012 | 55 | 35 | 3.7 |
From this capsule, it’s much harder to draw any conclusions.
Several little things could be distorting what is already a rough study:
- Most teams haven’t actually played 30 games yet this year; the average through Monday was 29 games. All teams have played at least 28 games, but I didn’t do a 28-game study because…
- Some of the closer changes have occurred recently, so eliminating even one or two games for those teams might exclude the very evidence I’m after.
- It may be that “saves per closer” is ineffective at tamping down all the random little variables that can lead to more than one closer being used even though no strategic change has been made.
All in all, while I hate to flush my work, it’s probably too early in the year to study this question. But if any of you want to pursue it, here’s the full data set:
[table id=48 /]
