Andrew Miller at Nick Markakis Money

As you understand and have presumably criticized, Nick Markakis has signed with the Braves, for four guaranteed years and a reported $44 million in guaranteed money. While the Braves know things we don’t, and while every team deserves some kind of benefit of the doubt, the signing seems puzzling, especially for a team that appears to be focusing more on contention down the road. We’ve written about trying to figure out why Markakis got so expensive. The numbers we have don’t exactly support the full commitment.

Later today, or maybe tomorrow, Andrew Miller is likely to sign with a team. Though he’s a free-agent reliever, he’s a virtual lock to get four guaranteed years, and if he doesn’t get $44 million, he’ll get almost that much. This is the state of the market, where Zach Duke gets three guaranteed years following a surprising one-year sample. However, while ordinarily you’d think you’d rather have an expensive right fielder than an expensive reliever, in this case the numbers are more supportive of Miller. Miller’s going to get Markakis money, and Miller is arguably more deserving.

Maybe you know how good Miller has been. Maybe you don’t. Maybe you understand that he’s been good, but maybe you don’t fully grasp the extent to which he’s dominated. This is really simple. Miller became a full-time reliever in 2012, after busting as a talented starter. Last season he was worth 2.3 WAR, or 2.0 RA9-WAR. Now, over the last three years, 171 different relievers have thrown at least 100 innings. Here is where Miller has ranked in a few categories:

2012 – 2014

• ERA-: 12
• FIP-: 5
• xFIP-: 5

Very impressive! But that also doesn’t quite capture Miller’s upward trajectory. Let’s look at last year, alone. There were 171 different relievers who threw at least 40 innings. Here’s where Miller ranked in the same categories:

2014, only

• ERA-: 19
• FIP-: 4
• xFIP-: 2

Still, that doesn’t quite show enough. Miller’s a lefty reliever, and he’s always been a lefty, and so in his first year as a reliever, he faced a lot of lefties. Unsurprisingly, he kept left-handed batters under control. But this chart shows three separate and individually impressive trends:

In his first year as a reliever, Miller averaged a ratio of two righties and three lefties. Last year he reversed that, as both Boston and Baltimore trusted him against any and all hitters, across the board. There were reasons for that, as shown by the other two lines. Miller has doubled his K-BB% against lefties since becoming a reliever. And he’s more than doubled his K-BB% against righties, shooting up from 12% to last year’s 30%. Usually, relievers have heavy platoon splits. That’s why they’re relievers in the first place. Miller transcended that, becoming less of a left-handed reliever, and more of a general elite reliever. With the elite ones, you don’t care who they might be facing at the plate.

It’s an explanation you seek. Miller has tweaked his delivery from time to time, and this .gif, courtesy of Texas Leaguers, shows you little adjustments he’s made to his arm slot:

Miller now releases the ball a little closer to the rubber, and from a little bit higher. The effect of this, and other things, has been improved command of the slider, which functions as Miller’s co-primary pitch. A heat-map .gif, from Baseball Savant:

You can see how, last year, the slider locations were a lot more focused. Just about every time, Miller wants to throw his slider either in the corner off the plate, or in the corner just on the plate. That’s exactly what he was able to do pretty consistently, and you can see the evidence in other numbers. Miller’s zone rate has stayed steady. His zone-contact rate has stayed steady. His out-of-zone-contact rate has dropped, almost entirely because of the slider. Between 2012 – 2013, Miller’s slider posted a 40% O-Contact%. Last year, 27%. That’s just about the whole story. Spotting the slider allowed Miller to unlock an extreme new level.

So not only does he have a three-year track record of success — he’s shown improvement, now coming off of a career year. There’s reason to believe Miller will be about as good as he just was, and that’s why he’s been able to ask for Jonathan Papelbon money. That’s also why he’s going to get it. He’s not Aroldis Chapman, but Aroldis Chapman isn’t available.

It shouldn’t be surprising that Miller has been most closely linked to huge-market teams. Though he has numbers on his side, it’s the bigger payrolls who can feel safest giving such a commitment to a reliever. The more a team has to spend, the less it has to worry about its payroll efficiency, so given the Yankees’ money and their opening in the bullpen, they’re an obvious match. The Red Sox would also be a match, and the Dodgers would be another match. The most curious rumor, hands down, has been the one linking Miller to the Astros.

That seems unlikely to come to fruition, but the existence of a rumor in the first place is odd, given that the Astros project to be not particularly good. Maybe it’s that the Astros realize they don’t have many commitments on the books, and they like what Miller would be able to bring to what’s been a lousy bullpen. But there’s also the chance the Astros would be looking to flip Miller in a midseason trade. There’s risk in that, but they’d only need to pitch Miller for a few months, so the risk would be fairly low. And prices go up during the year, in particular for relievers. Last summer, Miller, Joakim Soria, and Huston Street got moved for quality prospects, and if the Astros were to sign Miller, it might just be a move on the path toward acquiring more youth. Ultimately, though, Miller doesn’t seem Houston-bound. The Astros might opt to try the same sort of maneuver on a smaller scale with Sergio Romo.

Miller has a four-year contract coming, from someone. The contract will cover his age 30 – 33 seasons. Based on market rates, he’ll be paid to be worth around 5 – 6 wins. Over the last 30 years, 27 relievers have been worth 5+ WAR between 30 – 33, and 16 relievers have been worth 6+. Meanwhile, 61 relievers have been worth 5+ RA9-WAR, and 44 have been worth 6+. It’s true that relievers are less predictable than other players, but a lot of that has to do with the sample sizes. It’s easier for things to go wonky over 60-odd innings than over 150 innings or 500 plate appearances. Whoever signs Miller will be looking for 250 innings or so, and that’s a more predictable sample. He’s been an elite reliever, so he ought to remain around the level, barring injury. It’s always barring injury.

Miller was drafted one slot in front of Clayton Kershaw in 2006. At the end of 2011, he had a career ERA pushing 6. No longer can he be considered a total bust; now he’s just an unexpected kind of successful, and he’ll have an eight-figure contract before he turns 30. There is such a thing as a reliever better than Andrew Miller. There just aren’t very many such things.