In Monday’s column, I picked on Keith Foulke a bit, expressing the opinion that I didn’t think he’d sustain his performance throughout the life of his new four-year contract. Buried in that criticism was the following:

I haven’t looked too deeply at this yet, but I don’t think you can find a lot of relievers who stayed at Foulke’s recent level for a seven- or eight-year period. I think the best relievers in the modern era have short, high peaks before slipping. Mariano Rivera is the exception to this, but when you look around at the best relievers in baseball, by any standard, there just aren’t guys who are worth five or more WARP a year for most of a decade.

Just looking at some names, there are a lot of guys who had a brief monster peak, then settled in at a lower one…Troy Percival, Armando Benitez. Billy Wagner and Robb Nen probably have the best performance records aside from Rivera, but each missed a season due to injury, as did Hoffman. There’s just not a lot of history to support the idea that Foulke is going maintain his level of performance and his durability throughout the contract.

I decided to take a look at this, because while my gut tells me that relievers burn fast and bright, and that even the best ones don’t have much longevity, I don’t have a lot of hard evidence for the notion. I do, however, have a Sabermetric Baseball Encyclopedia and the BP player cards, which seemed like pretty good places to start.

I identified the relievers who, like Foulke, had generated the most value during their late 20s. Foulke has been worth 29.2 wins above replacement at ages 26-30, the third-highest WARP figure since 1980 of any relief pitcher in that five-year span. Let me just run the chart:

WARP1 for relievers in their age-26 through age-30 seasons

Pitcher           WARP1
Mariano Rivera     33.5
Trevor Hoffman     29.3
Keith Foulke       29.2
John Wetteland     28.0
Goose Gossage      27.3
Robb Nen           27.1
Lee Smith          25.7
Jeff Montgomery    25.1
Jesse Orosco       25.0
Armando Benitez    23.8
Tim Burke          23.7
John Franco        23.7
Bruce Sutter       23.3
Willie Hernandez   23.1
Billy Wagner       22.5
Jeff Reardon       22.0
Duane Ward         21.5
Troy Percival      21.4
Bryan Harvey       20.1

This is Foulke’s peer group: the guys who dominated games for five seasons up to age 30.

For the purposes of the study, however, I removed three people: Rich Gossage, Bruce Sutter, and Willie Hernandez. In looking at their usage patterns, it seems to me that including them in a comparison to Foulke would not be helpful. All three were more ace relievers than modern closers. I drew the line between those three and Jesse Orosco and Jeff Reardon for purposes of this exercise.

That leaves 15 pitchers other than Foulke. Here is what their average line looks like from ages 26 through 34. Note that all these averages include all players who have reached that particular age; missing a season counts as a zero. (Not yet being that age, however, doesn’t hurt the group, and accounts for the dwindling number of subjects in the latter age groups.)

Age     IP    ERA    K/9   K/BB   WARP
26    87.2   2.46   9.39   2.78   5.21
27    79.2   2.66   9.67   2.82   5.15
28    78.0   2.55   9.11   2.71   4.99
29    71.0   2.84   8.67   2.86   4.81
30    70.0   2.73   8.77   2.81   4.65
31    60.1   2.75   9.24   3.34   4.20  (14 pitchers)
32    52.0   2.89   8.81   3.22   3.42  (13 pitchers)
33    47.0   3.29   7.39   2.47   2.72  (13 pitchers)
34    34.2   3.25   7.38   2.49   2.14  (10 pitchers)

I think one of the most remarkable findings for me was the attrition rate of the game’s top relievers. Of the 15 pitchers included in this study, all of whom racked up at least 20 WARP from ages 26-30, four were out of the game by the age of 34. John Wetteland, Tim Burke, Duane Ward, and Bryan Harvey were all done just four seasons past their peak.

The truncated careers account for enough of the plummeting WARP averages to warrant looking at the data without them:

WARP figures for only pitchers throwing at least 15 innings in a season

Age 30: 4.99
Age 31: 4.97
Age 32: 4.05
Age 33: 3.53
Age 34: 3.57

Collectively, relievers who are dominant in their 20s do decline in their early 30s. Much of that risk is in the form of injury, rather than performance deterioration. The real break for the guys on the field seems to occur between 31 and 33, where the pitchers who had formerly been dominant step back to being just good, often at the same time they become very expensive. The chance of getting a star-caliber season from a top relief pitcher after they turn 30 is less than half of what it is before that.

WARP of 6.0 or better, by age:

26: 4 of 15
27: 5 of 15
28: 3 of 15
29: 4 of 15
30: 5 of 15
31: 3 of 14
32: 1 of 13
33: 2 of 13
34: 0 of 10

Factoring out the player-seasons yet to come, pitchers from 26-30 have put up six-win seasons 28% of the time. After age 30, this same player pool puts up six-win seasons 12% of the time.

What does this all mean for Keith Foulke? Before doing this research, I wrote in an e-mail that he had about 120-150 innings at or near his current level left in him. That would be a shade shy of two seasons, and is a notion completely supported by the career patterns of those who came before him. Moreover, the risk that he’ll get injured and be a $6 million albatross can’t be dismissed, given that not only did four pitchers never make 34 in a uniform, but that others who did, such as Trevor Hoffman, Robb Nen, and Billy Wagner, all lost a season along the way to injury.

What I take from this research is a reinforcement of the idea that there is no reliever in the game worth the risk of a long-term free-agent contract. They don’t age well, they miss entire seasons with too much frequency, and the salaries they command on the market make it nearly impossible for them to end up as a good value.