The Yankees Haven’t Had a Below-Average Bullpen In 20 Years

The Yankees have assembled one of the most intimidating bullpens imaginable. This is covering ground that’s been covered, but seriously, you’re not doing anything better with your time — think about this again. The Yankees have Aroldis Chapman, Andrew Miller, and Dellin Betances. Last year they threw 212 innings, with 347 strikeouts and a 1.66 ERA. The year before, they did the same thing. It would be a luxury to have just one of these pitchers, and this feels like something off of a fantasy roster. If the Yankees keep this group together, it’s going to help the starting rotation, and it could be an unhittable daily force in the playoffs.

As I wrote about the Chapman trade last week, I discussed how this should only continue an organizational pattern. Last year’s Yankee bullpen was outstanding, with Justin Wilson in place of Chapman. The Yankees have gotten used to strong bullpens, with Joe Girardi proving himself a skilled high-leverage manager. I didn’t think too much about this a few days ago, but now that I’ve taken the time to look back, it’s been ages since the Yankees had unreliable relief. According to how they’ve actually performed, the Yankees haven’t had a below-average bullpen in two decades.

Maybe that’s stunning. Or maybe it’s not! It’s up to your own interpretation, and of course, you know the Yankees had the whole career of Mariano Rivera. That’s the sort of thing that makes a difference. And on the biggest-picture scale, the Yankees haven’t had a below-average team since 1992, when they lost 86 times. The Yankees are the Yankees, with all the Yankees’ resources at their disposal. They’re supposed to be above-average. There’s not really a good excuse for them ever being below-average. I think that rings most true overall. I find it more astonishing when you look at one individual component. Like, say, a team’s relievers.

We host Win Expectancy information going back to 1974. So let’s examine that window, with the focus on the Yankees bullpen. The blue line shows the Yankees’ year-by-year bullpen WPA. The gray line shows the league-average bullpen WPA, per whatever number of games the Yankees played.


You see mostly good Yankees performances, with some dips in the early- or mid-90s. Through the first 22 years shown, the Yankees ranked third-best in baseball in bullpen WPA. But now, draw a line between 1995 and 1996. It was in 1996 that Mariano Rivera made a permanent switch to relief. Here’s the resulting plot:


20 years of bullpen performances, and 20 years of above-average bullpen performances. On a few occasions, the blue line threatens to dip below the gray line, but it never quite gets there. The differences have been consistently enormous since Girardi was hired for 2008 — the smallest difference has been +2.4 wins. But this isn’t just about Girardi. This is about the whole 20 years. The Yankee bullpen hasn’t struggled once. If you go with the simple-math expectation that a team has a 50% chance of having an above-average bullpen in a given year, then the chances of 20 consecutive years like this would be more slim than one in a million.

That is, if things were happening by chance. This isn’t by chance. The Yankees have had Rivera. They’ve had Girardi. They’ve had money! Over the two decades, the Yankee bullpen has averaged +5.8 WPA per year. The league has come in at +1.6, for a pro-Yankees annual difference of about +4.2 wins. The Yankees bullpen has dominated, and it’s dominated even the second-best bullpen over the selected window.


That’s 20 years of team bullpen performances, combined. Not quite for Arizona and Tampa Bay, but, anyway. The Yankees are in first by a huge margin over the Angels, who then hold a relatively slim lead over the Twins and Rangers. The gap between first and second is 38.5 wins. The gap between second and thirteenth is 38.9 wins. Over 20 years, Yankees bullpens have been worth something like 123 more wins than Cubs bullpens. That doesn’t have to mean much of anything for the future. The Cubs, naturally, have turned a few pages. But looking back, that’s more than six wins a year from relievers alone. Cubs fans would probably tell you it’s felt like it.

It’s no secret the Yankees were blessed by Rivera’s development. Nearly 2,800 pitchers have relieved since 1996. Rivera ranks first in WPA, at +58.0. In second place you find Joe Nathan, all the way down at +30.2. Rivera leads the second-place Yankees reliever — David Robertson — by a margin of +58.0 to +11.9. It only makes sense the Yankees would look amazing when you select for Rivera’s career, because he’s probably the best reliever the game has ever seen.

But now, for fun, take Rivera away. Just delete him. The 20-year Yankees bullpen drops from +116 WPA to +58. That would still have them ranked fourth in baseball over the span, behind the Angels, Twins, and Rangers. Reality would be more complicated than this — that isn’t a measure of what the Yankees would’ve done if they never had Rivera on the team. But this just goes to show that, while Rivera has been the single most important reliever, the overall success has been a function of the group. You could say it’s been half Rivera and half everybody else. And since Rivera left, it’s not like the Yankees have struggled.

Nor should they in 2016. To this point, they’ve put together 20 consecutive years of above-average bullpen work. It would be an absolute shocker if they didn’t make it to 21. The Yankees are always supposed to be successful. They want to be better than average in every department in every season. They don’t always get their way. In the bullpen, though, they’ve had little to complain about. Yeah, Rivera was something special. One player alone can’t explain two decades.

Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.

Newest Most Voted
Inline Feedbacks
View all comments
Jetsy Extrano
8 years ago

The historical graph of league average bullpen WPA is really something! It looks like it trends gradually up from 0.5 wins in the olden days, to the current 2 wins. I mean we know bullpens have changed, but there it is.

You already wrote about this, didn’t you?

Doug Lampertmember
8 years ago
Reply to  Jetsy Extrano

The historical graph is interesting because I typically assume that current reliever usage is inefficient and that closers should instead be used as firemen.

And while this doesn’t actually prove that I’m wrong about closer usage, it does at least show that the current pattern still seems to be an improvement on what I grew up with.

Managers have larger bullpens and more short guys, and it is in fact producing wins from relievers. I’d be curious as to how bench player WPA works out over the same span (the extra relievers come at the expense of bats on the bench), but that’s probably harder data to assemble and I actually doubt that the loss of bench players hurts anything like as much as the added arms are helping.

8 years ago
Reply to  Doug Lampert

Improving for sure… though relievers themselves have generally gotten better on the whole…

Also, don’t confuse inefficient and ineffective. Versus a random distribution of RP innings, your dad’s old “save the closer for the 9th only if we’re winning” does get you a good way along toward an efficient use of a bullpen. Obviously no where close to fully efficient, but it’s not value-destroying…

Jetsy Extrano
8 years ago
Reply to  Doug Lampert

Yeah, the current bullpen pattern can be less efficient while still yielding more wins, if the investment — in roster spots and other opportunity costs — has gone up enough.

They may be losing wins that used to come from the bench as you say, or from platooning the lineup, or from relief pitchers who could have been average starters… That would be interesting to see.

8 years ago
Reply to  Jetsy Extrano

The 0.5 WPA was one blip in a 20 year curve. The first point in 1996 closer to 2.0

WPA is a counting stat and RP’ers pitch far more innings than 20 years ago so you would expect an increase, and there has been, just not as much as you believe

Jetsy Extrano
8 years ago
Reply to  Paul22

Look at the graph back to 1974.