Monday, March 21, 2011

Thoughts on Bullpens: Relievers, and Why They Don't Matter

Over the course of the last decade it's been pretty common to hear stat-heads talk about how the closers role is overblown. It's not a difficult case to make. If I were to take a quick survey of 100 random baseball fans and ask them who was the best closer in baseball, they'd probably name Mariano Rivera. They'd be right. Even last year as his age finally began to show, and his strikeout rates dipped, Rivera remained fantastic at closing out games.

Yet even with a sparkling 1.80 ERA, Rivera was just the 6th most effective reliever in the AL last year. Over the past three seasons Rivera has led baseball in ERA with a 1.64 mark, but six others have also had ERAs under 2.00. So while Rivera may be the most dominant reliever in baseball, it's not by very much. In fact the difference between a 1.64 ERA and a 2.00 ERA over 180 innings is very small - 32.8 runs vs 40.0 runs. 7.2 runs in total. Or on average, 2.4 runs per year.

So yeah, Mariano Rivera is amazing. No doubt. But in terms of overall effect on a team? Not really any different that a handful of other guys.

This is true on a larger scale as well. While we are inundated on television and radio by announcers and analysts who seem to take great joy in extolling the virtue of the closer, or the importance of a teams bullpen, the truth is this: they don't matter.

Well, not much anyways.

I understand a few things here. One, I'm not saying something that hasn't been said by a couple hundred other writers. Two, for a lot of folks, it's baseball sacrilege. However, it's something that I think still needs to be said.

Consider if you will - the best bullpen in the AL last year belonged to Tampa Bay. That's not a huge surprise, they had Rafael Soriano, Joaquin Benoit (now with our very own Tigers), Grant Balfour, and Randy Choate amongst others. The fact that all of those pitchers are now employed by other teams is one of the reasons why many are expecting a significant drop in the W/L column for the Rays in 2011.

Just today, Jon Morosi, a quality national baseball writer with a long standing tie to the Tigers after covering them as a beat reporter used the fact that the Twins had lost many from their bullpen, including Matt Guerrier, Jon Rauch, Jesse Crain, and Brian Fuentes as a reason for concern about the Twins chance in 2011, saying "the Twins simply lost too many important relievers." Morosi would go on to say, "The rotation is a concern, too" in regard to the Twins. Never mind the fact that the same rotation was good enough to be best in the AL Central last year - with no one having made any meaningful improvements. But that's a story for another day.

For now let's focus on the part relevant to the subject matter. Have the Twins really lost too many important relievers to contend? First, lets set some background. The 2010 Twins came into the season with a strong, but not great group fronted by Joe Nathan. In 2009, essentially that exact same group was 10th bet in baseball with a 3.87 ERA. The Nathan went down.

The national media did it's usual dance. The Twins had a huge hole to fill, how would they do so? Could they? Was the Twins season hanging in the balance? We know how the story played out of course. A Nathan-less Twins bullpen would go on to finish 8th best in baseball with a 3.49 ERA.

Not only that, an unproven reliever, Jon Rauch - gasp! - managed to step into the role seamlessly. Stunning, I'm sure. Yet there it is, plain as day.

So how could this have happened? How could a unit that was 10th best one year be even better the next while losing it's best player? Well, part of it was luck. As a team the Twins outperformed their combined relief FIP of 3.49 by 0.42 runs. But even when you look at it by FIP, the Twins were 13th best in baseball, solidly average.

The bigger part of the Twins 2010 success - or stunning lack of collapse as some TV/radio/print analysts might have you believe - was the ugly truth. Bullpens don't matter.

It's true.

It's true whether you want to look at it from a saves perspective (closers aren't mentally tougher than other relievers) an ERA, or a WAR perspective. All cases show essentially the same thing - the difference between the best bullpens and the worst bullpens simply isn't that great. Don't believe me? Good, lets look at the numbers.

As stated earlier, the American Leagues best bullpen in 2010 belonged to the Tampa Bay Rays. Their staff combined to throw 454 innings of 3.33 ERA baseball, allowing a total of 168 runs (or 1.037R/G). The Royals came in dead last. Their bullpen threw 496.2 innings with a 4.46 ERA, allowing a total of 246 runs (1.518R/G).

Clearly the Rays bullpen benefited from having a better starting staff as well, and needed to throw 42.2 innings less. On average, AL bullpens threw 463 innings, allowing 200 earned runs, for a 3.89 ERA.

So the first thing we need to do to accurately compare the two bullpens, is look at runs allowed per innings (ERA) and balance that out over the league average. When we do that, the runs allowed number becomes 171.33 for the Rays, and 229.33 for the Royals.

The total spread is just 58 runs. Fifty eight runs, that sounds like a lot, doesn't it? But how much is it really? Well, in practical terms, it's about one run every three games. Could that run be the difference between a win and a loss. It could, but probably not. Most games are decided by more than one run after all. In sabermetric terms, the difference is greater - 5.8 wins. Six games after all is a lot, especially when a team is in a playoff race. Of course, it's important to remember that were discussing the difference between the best bullpens and the worst bullpens. Most teams by rule, are neither.

What about the Twins then? Are they really going to be that much worse off in 2011? Let's do some math, of the players the Twins lost from 2010, they received a total of 2.2 WAR - 1.9 of which came from the combination of Jon Rauch (1.1) and Jesse Crain (0.8). The other four departing relievers combined for the other 0.3 WAR.

So in other words, the Twins would need to find some combination of four relievers to go along with the returning trio of Joe Nathan, Matt Capps, and Jose Mijares to who could make that up. To make the ease of making up 2.5 WAR between four players even more laughably easy, let's assume the Twins don't get any more than the 0.6 WAR they got from Capps in just 27 innings (he'd be worth about 1.5-2.0 over a full season). Let's also assume they get precisely 0.0 WAR from Nathan - who has average 2.52 WAR by himself the last two years.

Yes, lets do Morosi and all the other national writers predicting gloom and doom for teams like the Twins and Rays a favor by ignoring all the obvious stuff, and call it all irrelevant. Let's say the Twins get NOTHING more from either of those guys.

We know one of the players who will make the Twins bullpen in 2011 will be whichever starter doesn't make the rotation. As of now, that looks like it'll be Kevin Slowey. Over the past three seasons Slowey has been worth 2.2 WAR per year while averaging 133.2 innings. Assuming he throws about half that many out of the pen (68IP), he'd be worth about half as much in terms of WAR - or 1.1.

That leaves just 1.1 WAR left for the three other pitchers to make up. Now let's get really crazy and say that the other three pitchers are so horrendous that they aren't even replacement level, and end up worth -0.9 WAR, for a combined loss of 2.0 WAR.

That mean that in what amounts to essentially an absolute worst case scenario, a team that won 94 games could stand to lose 2 more games because of it's bullpen departures. IF everything goes wrong. IF Joe Nathan gives the team squat. IF Matt Capps has a bad year. IF they actually get negative production from their other pitchers.

If everything goes wrong, the Twins might lose a couple more games.

I shouldn't need to explain to anyone how unlikely it is that all of these things actually occur. In truth, the Twins 2011 bullpen might very well be worse than it's 2010 iteration - maybe even by the 20 runs necessary to create that full 2.0 WAR, worst-case-scenario difference. But it isn't likely.

In deciding to do this piece using Morosi's article as a backdrop, I'm probably being a bit unfair to him. It's not just Morosi who does this after all. The bullpen as a valuable tool, and relievers as irreplaceable commodities meme has gotten beaten to death for 20 years by the mainstream media.

In choosing to add my voice to the chorus of people who fight back against such blatant negligence, I'm taking a risk of sorts. I understand that those who don't buy into advanced metrics will quite simply be dismissive of such obvious evidence. But whether you choose to believe in it or not, no reliever, not even Mariano Rivera is so much better than someone else that he can't be replaced rather easily. Nor is any bullpen, no matter how decimated by injury or departure, so far gone that it can't be rather easily fixed. 

Corey Ettinger is a Senior Writer for Baseball Digest as well as a proud contributor to both,, and He also provides extensive analysis of the American League Central Division at his own blog, AL Central In Focus. Be sure to follow me on Twitter @Coreyettinger for the latest updates, random thoughts and general tomfoolery.


  1. I agree that bullpens (especially closers) are overrated, but you're going way too far the other way, and completely dismissing their value. You mention making up 2.5 WAR as being laughably easy, and even say 6 WAR isn't that much.

    Miguel Cabrera has averaged 4.64 WAR in his 7 full seasons. How easily replaceable to do you think he is? Mariano Rivera has averaged 3.52 WAR over the past 15 years.

    I'd have expected that difference to be larger, but you make it sound like they'd be in different stratospheres.

  2. It's between seven players though, not one. In total it requires less than one WAR per player.

  3. But 1 WAR relievers don't grow on trees. Valverde was the Tiger's only reliever above 1 WAR last year according to baseball reference.

    And to again show that elite closers aren't as unimportant as you think, Verlander's average WAR - 3.78. Not far above what Rivera averages.

    I think the elite closers still have elite value, it's more your average closer that tends to be overrated and overpaid. Like Valverde's 7 mil contract only gave us 1.2 WAR last year, and on top of that we had to give up a first round pick to sign him.

  4. Ok, lets try this another route, because the initial explanation didn't seem to work. We'll do bullet points.

    The Twins had a "good" bullpen in 2010.

    They had one reliever with a WAR over 1.0

    They lost a total of 2.5 WAR.

    Those two wins are what are being described as a severe loss.

    It isn't because....

    2.5 WAR is easily replaced.

    Especially when divied up between five players.