Brad Miller is showing his flaws with more playing time

A little more than three months ago, I published a piece called “Brad Miller is more valuable than you might think.” I had this to say about Miller’s performance at the time:

Over the past three seasons, Brad Miller has slashed .255/.345/.497 with a 123 wRC+ in 415 PA. (This includes his time with Cleveland, St. Louis, and Philadelphia.) A 123 wRC+ puts Miller in the top 20% of all hitters during that time span (min. 400 PA). For those unfamiliar with wRC+, it is a statistic that attempts to quantify a player’s total offensive value and compare it to league average. A 123 wRC+ means that Brad Miller has created runs for his team at a rate 23% better than a league-average player, and 5-6% better than Rhys Hoskins (118 wRC+) and J.T. Realmuto (117 wRC+). That’s really good.

Just over a week after I sent those words out into world, Brad Miller went ice cold. From May 25 until August 29, Brad Miller hit just .160/.268/.355 with a 67 wRC+. He turned from an undervalued offensive gem into a pumpkin (or, in terms of wRC+, he went from Juan Soto to Elvis Andrus) seemingly overnight. His season slash line dropped to .217/.307/.421 with a 95 wRC+ as of August 29.

I read many an article (and angry tweet) over the past few weeks that touched on Brad Miller’s struggles. Phillies fans weren’t happy with Miller’s performance, to say the least.

But then – to further complicate things – he could still sometimes put on an absolute show. Remember July 8, when he hit three home runs in one game? Or how about last night when he went 3-for-3 with a home run and two walks and was named Player of the Game on the TV broadcast? With his incredible showing last night, he raised his 2021 wRC+ from 95 to 103.

So who is the real Brad Miller? Is he the player that posted a 130 wRC+ through 163 games from the beginning of 2019 until May 24, 2021? Or is he the player who’s hit below the Mendoza-line since?

Unsurprisingly, the truth is that the real Brad Miller lies somewhere in the middle.

For one thing, no one should expect Miller to be a 130 wRC+ hitter over a full season. Otherwise, he wouldn’t be a Phillies bench player on a one-year, $3.5 million contract. However, he has also been the victim of some bad luck on balls in play recently. Miller had a .188 BABIP in 73 games from May 25 to August 29. His career average BABIP is .288. His Statcast numbers further support the argument that he ran into some bad luck. His expected wOBA during that slump was well above league average, while his actual wOBA was a pathetic .271.

But bad luck can’t explain it all. Not when Miller had been so bad over so many games. No, the root cause of Miller’s mediocre performance this season is more than just some bad luck. The bigger issue is that the Phillies have been relying on Miller far too much recently, and therefore his biggest weakness has been exposed: facing left-handed pitching.

I’m not going to go into Miller’s history of struggles against LHP here. I wrote about that all the way back in April. But let me just share his 2021 splits.

• Brad Miller in 2021 vs RHP: .246/.356/.476, .832 OPS, 123 wRC+
• Brad Miller in 2021 vs LHP: .171/.205/.343, .548 OPS, 44 wRC+

Against RHP, Miller is still having a strong season, but it’s clear that his numbers against LHP are tanking his overall season statistics. The reason that’s happening is because Miller has had far more plate appearances against left-handers than usual this year.

Brad Miller has had 292 PA this year – 219 against RHP and 73 against LHP. That means 25% of his PA have come against southpaws, which is pretty close to the 2021 NL average of 27.5%. But when Brad Miller is being deployed correctly, he rarely faces left-handed pitching. In 2020, he only faced LHP in 20 out of 171 PA (11.7%). In 2019, he only faced LHP in 22 out of 170 PA (12.9%). That’s just 12.3% of his PA against lefties from 2019-2020. In other words, he has more than doubled his rate of PA against left-handers this season. No wonder his numbers don’t look as good as they used to.

So what if Miller hadn’t had so many plate appearances against lefties? How would his numbers look if he was only facing lefties 12.3% of the time, like in 2019/2020? The answer is pretty impressive.

.236/.337/.458, .795 OPS, 119 wRC+

Obviously, this is just an estimation. There’s no official way to calculate something like this, so I’ll briefly explain my methodology.

I kept Miller’s numbers against right-handers constant. He has had 219 PA against RHP this year, and I didn’t change that. I didn’t add plate appearances against RHP to make up for the ones I was taking away against LHP.

Instead, I figured out how many plate appearances against LHP he would need in order to make up 12.3% of his total PA. The answer was 31 (technically that’s 12.4%, but it’s the closest we could get). 31 PA vs lefties plus 219 PA vs righties adds up to 250 total PA in this theoretical season.

Miller’s BB% against LHP is 4.1%, which means in 31 PA he’d walk 1.271 times (let’s just pretend that’s possible for the sake of these calculations.) That means he’d have 29.729 at-bats. Presuming he keeps the same BA/OBP/SLG, he’d have 5.083659 hits, 6.355 times on base, and 10.197047 total bases.

I added those totals to his actual numbers against RHP (46 hits, 78 times on base, 89 total bases) and recalculated the rate stats. That’s how we get to these numbers:

.236/.337/.458, .795 OPS, 119 wRC+

Those stats look very similar to the ones I wrote about back on May 18 when I was praising Miller’s offense. And with his 2021 BABIP back up to .288 after his 3-for-3 showing on Monday night, we can trust that these numbers are reflective of Miller’s true talent level. Thus, Brad Miller isn’t having a markedly worse season in 2021, he’s just been asked to take on a bigger role than he’s capable of.

So the next time Brad Miller steps up to the plate, make sure to remember who he is. If a right-handed pitcher is on the mound, he’s arguably the second-best hitter in the lineup. But if he’s facing a left-hander, keep your expectations low. Very low.