clock menu more-arrow no yes

Filed under:

Will the Real Andrew Miller Please Stand Up?

New, 63 comments

The lefty’s first season in St. Louis was a roller coaster

MLB: Pittsburgh Pirates at St. Louis Cardinals Billy Hurst-USA TODAY Sports

Andrew Miller’s first season in St. Louis had quite a few ups and downs, and it left him with disappointing numbers at the end of the season. The lanky lefty pitched to a 4.45 ERA, 5.19 FIP, and his -0.86 WPA (win probability added) was second worst on the team. His FIP was 12th worst among 158 qualified relievers and his WPA was bottom 30 as well. With all of that in mind, it has become fashionable to look only at the FIP amassed in 54.2 innings and label him a busted reliever. There’s a lot more nuance to that story, though.

Reimagining 2019

Let’s reimagine the 2019 Cardinals season. In this new scenario, the Cardinals didn’t sign Andrew Miller after the 2018 season. Instead, they signed a minor league free agent named Larry Lefty who, as you likely guessed, is a left-handed reliever. He shows up to spring training on a non-roster invite and takes the camp by storm. Larry’s fastball has above average lefty velocity and his slider is a death knell for opposing hitters when it’s working. Larry is so impressive when he breaks camp that he’s the top lefty called on by Mike Shildt as the season begins. The plan even calls for him to see plenty of high leverage work to start the season.

The season arrives and Larry’s impressive March innings have evaporated. Suddenly, he’s routinely getting his nipples lit up like a Christmas tree. He’s one of the worst relievers in baseball. It’s especially troublesome early in the year when he’s getting high leverage innings, but eventually Mike Shildt determines that Larry needs to be moved to something with less pressure.

This goes on until late June when the team has seen enough and Larry Lefty is designated for assignment. The final week of June is spent with lefty Quentin Quadier (pronounce it like you’re French, if it helps with the soul-shredding wordplay) soaking up some innings and getting predictably hammered as John Mozeliak works the phones seeking a replacement. He finds Sammy Southpaw sitting on waivers. His presence, along with Larry Lefty, Quentin Quadier, Giovanny Gallegos, Miles Mikolas, and Mike Mayers gives the Cardinals the league lead in alliteration.

Sammy Southpaw dominates the league. For the rest of the season, his FIP is top 20 among relievers. His xFIP is third best. He’s an absolute force and his performance, in combination with several other relievers, gives the Cardinals one of the very best bullpens in the game. He even shows up in October and pitches some effective, valuable innings in the NLDS and NLCS.

Here’s how the numbers break down on the season for our fictional trio.

Three Fictional Cardinal Lefties

Category S. Southpaw L. Lefty Q. Quadier
Category S. Southpaw L. Lefty Q. Quadier
IP 25.0 26.1 3.1
H 19 21 5
HR 3 5 3
K% 40.6% 23.5% 10.0%
BB% 7.9% 13.0% 20.0%
FIP 2.81 5.68 19.11
xFIP 2.29 5.63 10.99

My scenario here is obviously Andrew Miller, with his season split in three, albeit in different order. Quentin Quadier was Miller’s brutal first nine days of the season. After that, he settled in for a dominant stretch through July 5th as Sammy Southpaw. Finally, the dreaded Larry Lefty section of his season was July 6th until the end of the year.

If you’d rather see it in graph form, here is Miller’s 2019 using rolling 15 game averages from Fangraphs. One graph displays K and BB% and the other shows FIP and xFIP:

The narrative about Andrew Miller as an avoid-at-all-costs reliever isn’t exactly true. For three months, he was everything the Cardinals hoped for when they signed him in December 2018. It wasn’t totally necessary because the team also received outstanding contributions from many other relievers in the first half of 2019. The presence of other highly effective relievers coupled with Miller’s dreadful first nine days dampened the perception of his first half. But make no mistake- he was a force.

Or... was he? Let’s bring another table to the... er... table. This time, you’ll see it split up into three buckets- his first half, his second half, and his glorious stretch from April 7th through July 5th. In other words, the first category in the table is just his first half with the dreadful first nine days removed.

Andrew Miller, 2019

Category 4/7-7/5 1st Half 2nd Half
Category 4/7-7/5 1st Half 2nd Half
K% 40.6% 35.5% 23.5%
BB% 7.9% 9.9% 13.0%
K-BB% 32.7% 25.6% 10.4%
HR/FB 23.1% 31.60% 15.4%
BABIP 0.348 0.327 0.246
ERA- 68 90 121
FIP- 65 109 131
FIP 2.81 4.73 5.68
EV 87.5 89.2 86.5
Barrel/BIP 8.2% 11.5% 5.7%
xwOBA 0.283 0.321 0.294
xwoBACON 0.427 0.442 0.309

First of all, compare and contrast the difference between the first half and the April 7 through July 5 category. It’s crazy how much 3.1 bad innings can sink a reliever’s numbers. If stripping out those games feels too cherrypick-y to you, feel free to stick to the two halves and neglect the April 7th through July 5th category.

In the first half, at least during the dominant stretch, he was pretty close to vintage Miller in a lot of ways. His K-BB% from 2013 through 2017 was 33.7%. The difference is the HR/FB, which was 12.1% in his five year run of Classic Miller but nearly doubled in his dominant stretch last season. We’re dealing with such small samples that we’re essentially talking about one to two additional homeruns. That’s significant and it also explains the difference in FIP (1.96 vintage compared to 2.81 during his dominant Cardinal run). However, we’re still talking about a very good 2.81 FIP with an even lower 2.29 xFIP after we normalize his homerun rates.

What’s fascinating is the trade off between halves. In the second half, he cut his barrel percentage allowed in half, cut his exit velocity by nearly three miles per hour, and his expected wOBA on contact shriveled. It’s less pronounced if we omit his first nine days of the first half, but the effect is still there. The deal with the devil to achieve those second half results, though, was a loss of 12 percent in his K rate and a 3.1 percent increase in his walk rate. Whether through luck or otherwise, he did a much better job of managing damage on contact, but gave up so much more contact that it far outweighed the benefits.

In the grand scheme, Miller needs to find a mix of the two halves, preferably more first half than second. A pitcher with his first half K and BB rates and second half EV, barrels, and xwOBACON allowed would be an absolute monster. That he was so effective for half of the season should tell you that he still has the tools to be something close to what he once was.

How he gets there is a subject for another day, though Michael Augustine had some great thoughts about it- particuarly about Miller’s slider- not too long ago. However he ends up there, a more consistent Andrew Miller closer to his dominant half of the season would provide a major lift to the 2020 Cardinals.