February 2, 2016

Fifth Column

What Would an MVP Reliever Look Like?

by Michael Baumann

the archives are now free. All Baseball Prospectus Premium and Fantasy articles more than a year old are now free as a thank you to the entire Internet for making our work possible. Not a subscriber? Get exclusive content like this delivered hot to your inbox every weekday. Click here for more information on Baseball Prospectus subscriptions or use the buttons to the right to subscribe and get instant access to the best baseball content on the web.

Subscribe for $4.95 per month Recurring subscription - cancel anytime. Subscribe for $39.95 for one year a 33% savings over the monthly price! Purchase a $39.95 gift subscription a 33% savings over the monthly price! Already a subscriber? Click here and use the blue login bar to log in.

In last week’s listener email episode of Effectively Wild, Ben and Sam didn’t answer the interesting half of a question, so I’m going to do it for them.

Following up on a previous episode that discussed how Josh Donaldson was traded before his MVP campaign last year in a previous episode, an emailer mentioned that the last person to be traded the year before an MVP season was Willie Hernandez, who was traded from the Cubs to the Phillies in May 1983, then from the Phillies to the Tigers in March 1984[1] before winning the Cy Young and the MVP that year. The question, after the setup, was this: “Do you think this will ever happen again—a full-time relief pitcher winning the Cy Young and the MVP in the same season? If so, what sort of season would it take? Would it be possible to see this sort of player coming, like being drafted first overall, or would they inevitably be a fluke conversion, a la Mariano Rivera or Wade Davis?”

Ben started by dismissing a potential reliever MVP, with a caveat: “barring a big change in reliever usage, where guys go back to throwing 100 innings a year,” and then pivoted toward a discussion of the sabermetric perception of relief pitchers that dictated the rest of the course of the conversation. It’s an interesting conversation about how star relievers don’t stand out the way they used to, and if you’re in the mood, I recommend dialing up around the 8:25 mark in the podcast, or even all of Episode 805, for more on that fork of the conversation.

But I want to take Ben’s caveat—“barring a big change in reliever usage” and chase that rabbit for a while.

This topic, like so many others, stuck with me after I first encountered it in The New Bill James Historical Baseball Abstract. James decried what he called “the Robb Nen” pattern of relief ace usage, because “If you use your relief ace to save a three-run lead in the ninth inning, you’ll win that game 99% of the time. If you don’t use your ace in that situation, you’ll win 98% of the time … While I don’t mean to dismiss the intangible costs of blowing a three-run lead, an average team would win 97% of those games if they brought in Bryan Rekar in that situation.”

That is, of course, nothing new to anyone who’s followed baseball in the past 15 years, and I quote James directly only because he states the case against pitching to the save rule so eloquently, and with such vivid and visceral use of Bryan Rekar imagery.

James then endeavors to design an optimal reliever usage pattern, which he simulates by computer. A closer used for one-inning saves, James concludes, would win an extra game for every 4.64 runs he saves. However, a relief pitcher brought in for one inning at a time, either for a three-out save or if the score is within two runs in either direction going into the bottom of the eighth or later, would earn a win for every 2.51 runs he saves. Ultimately, James decided the best way to use a relief ace would be when allowing one run would make a difference: For two innings at a time when the game is either tied or when the relief ace’s team is up by a run after seven innings, or for one inning at a time in other late-and close situations when he hasn’t pitched in a couple days. James says a reliever used in this pattern “would not save 50 games, but he could win 20, and he might win 30.”

After 15 years, the numbers probably don’t match up exactly with today’s game, but the theory still works—you’d want your best relief pitcher to throw more than the 65-70 innings a Wade Davis gets nowadays, and rather than waiting for three-run leads to protect in the ninth, have him pitch more in tie games and earlier in the game, if practicable.

The right relief ace could probably throw 120 innings in a season, maybe more, but not the way relievers are used now. Today’s setup men get up and down, and pitch on back-to-back days, and warm up without entering the game. All of that works if you’re pitching one inning at a time, but it probably wouldn’t hold up long-term for two innings at a time. We know that pitchers can throw 75 one-inning games or 30-35 games of five to seven innings, on average, and that starting pitchers can come into games on their throw days without particular risk, at least in the short term. So could a pitcher throw every other game, more or less, for up to two innings, without breaking down? Call it 60-70 appearances and 110-140 innings?

Well, Mariano Rivera appeared 61 times for 107 2/3 innings in a setup role in 1996. Scott Williamson appeared 62 times for 93 1/3 innings in his Rookie of the Year campaign in 1999. With proper caution (i.e. an “if he warms up, he comes in” rule or something like it), it’s not out of the question. College teams do something like this already—if the game is close even after six innings, in comes the closer to finish it off, and thanks to the sparse collegiate schedule, a closer might throw 100 pitches a week over three outings, but they’ll be spaced out to Friday, Sunday and Tuesday. Now, college closers absolutely break down from overuse, even in a four- or five-month season, but they’re also generally a lesser quality of athlete, compared to pro closers, and college coaches have no incentive to preserve their arms past graduation, which is not the case in the big leagues.

In all likelihood, relief pitchers who could fill this role are uncommon—there might not be 30 of them in all of baseball. Well, there probably are, but if there’s a pitcher with good enough stuff to pitch high-leverage innings and good enough durability to throw 120 of them, he’s not going to the bullpen.

Which brings up the question of why a team would invest in such a player, since apparently guys with 95 mph heat and more devastating two-plane break than the DC-10 fleet in the early 1970s grow on trees. That’s probably the real obstacle to a plan like this—it’s probably more trouble to manage Wade Davis to a 120-inning, ultra-high-leverage workload than it is to have Wade Davis pitch one inning at a time and develop Kelvin Herrera and Greg Holland, and resurrect Ryan Madson, with the same effect.

But the reason you’d be tempted to sacrifice a potential starting pitcher for a relief ace, even one who pitches 120 innings a year, is the same way such a pitcher could stake a legitimate empirical claim to the MVP.

Even Mike Marshall, in his ridiculous Cy Young campaign of 1974, in which he appeared 106 times and threw 208 1/3 innings, was only a 3.7 WARP player. Dellin Betances, who’s pitched as much as any reliever in the past two years, topped out at 3.6 WARP over 90 innings in 2014.

Even splitting the difference between Marshall’s quantity and Betances’s quality, let’s say the reasonable ceiling for WARP for a 120-inning reliever is around 4. That’s not even close to an MVP-quality season, but not only is WARP not the only metric around, it does not capture the reason for creating such a role in the first place: A relief ace pitches fewer innings than a starter, but the innings he does pitch are generally more important.

In the expansion era, 14 relievers have put up seasons of 6 WPA or better, a feat matched by only 41 starting pitchers in the same time. The top relief season among those belonged to none other than Willie Hernandez, the selfsame 1984 AL MVP who inspired this entire conversation in the first place.[2] Hernandez threw 140 1/3 innings with a 2.13 DRA, good for 4.5 WARP, but with a WPA of 8.65, the second-highest mark of any pitcher in any role since 1961, on par with Steve Carlton’s 1972, Pedro Martinez’s 2000 or Bob Gibson’s 1968.

There’s not much new ground to break in the role of context-dependent stats in award discussions, but while a cleanup hitter who drives in 140 runs arrives at that total because of circumstances partially outside his control, the whole purpose of revamping the relief ace role is entirely about context, and making sure that the team’s best reliever pitches as many innings at as high a leverage as possible, which makes context-dependent stats relevant in this case in a way they wouldn’t be for a position player, or even a starting pitcher, whose job is to perform when his turn is up, whether the game or season is on the line or not.

So could a modern closer win the Cy Young and MVP? Probably not. Can a full-time reliever? Absolutely. And we’ve already got the blueprint.