Miles Mikolas causes trouble with his curve



On Tuesday night, after weeks of reading everyone’s year-in-review post and reflecting back on 2019, we celebrated the new year and turned the calendar to 2020. And while the new decade has officially started, I’d venture to guess most of us would admit that December 30th, 2019 and January 2nd, 2020 aren’t really all that different. Sure, we flipped the calendar, but life didn’t completely flip with it. The new year is an arbitrary point in time made meaningful because of the way our culture has chosen to group time, not because it’s the best way to view our experiences.


We view baseball seasons in a similar way, from opening day (happy new year!) to the conclusion of the world series. It’s easy to pull up a player page and make a declaration about how that player’s season went. Take Miles Mikolas from 2018 to 2019. After leading the NL in wins in 2018, he led the NL in losses in 2019. His ERA rocketed up by more than a run, in large part because he gave up a ton of home runs – 27, to be exact, which tied for the 11th most allowed by an NL pitcher. He dropped to 184 innings pitched after (barely) eclipsing 200 the year before. His velocity declined across every pitch. Overall, his WAR dropped from 4.2 to 2.5.


Years don’t happen all at once though, and neither do baseball seasons. Miles Mikolas made 32 starts in 2019. Over the first eighteen, he was a below average pitcher. Over the last fourteen, he was arguably back in the top twenty starters in the NL.


Did I still pick an arbitrary point in time to divide his season? You bet (hello, All-Star break). Why? Because for some reason today, I found myself looking at this chart, which gets really interesting around mid-July 2019:

There are a few things to see here (and sorry for the mess). First, Mikolas really mixes up his pitch selection, as evidenced by the lines clustering together between 20% and 30%. Secondly, look at the trends that start to emerge in July 2019:

  1. After two months spent moving away from his curveball, Mikolas increasingly leaned on the pitch throughout the second half of 2019.

  2. He stopped throwing so many sinkers, replacing them with more curveballs and fourseamers.

When I saw that chart with the context of his sub-par performance in the early parts of the 2019 season, I expected to find that Mikolas’ had an effective curve in the first half that he didn’t use enough and a bad sinker that he used too often. Sometimes, pitching is simple: throw your best pitches more often.


Instead, I found that the sinker was his best pitch in the first half (according to Pitch Info’s run values) and that his curveball, up until the All Star break, had been a below average pitch. In the second half, his sinker was bad and his curveball was good. I didn’t find much any meaningful changes in either pitch’s velocity, movement, or spin from one half to the other to explain the differences – the season just flipped to the second half and Mikolas’ pitches flipped with it. Sometimes, pitching isn't simple.

With my original lede at a dead end, I started thinking about whether Mikolas’ pitch selection could inform what he thought about his own repertoire. Looking at his pitch selection in the playoffs might reveal if he felt something had changed – with higher stakes, he might be inclined to throw his best pitches more often. And in those two playoff starts, he ditched his rarely used changeup and traded out a handful of sinkers for sliders and fourseamers. Not too surprising, but nothing to support that his curveball became a go-to pitch.


What about his relief appearance? Remember, he tossed a 1-2-3 tenth in game four against Atlanta. In that inning, he threw ten pitches including six curveballs, three fourseamers, and a sinker. Six curveballs to Brian McCann, Nick Markakis, and Rafael Ortega. Markakis and Ortega are each equally competent at hitting curveballs compared to sliders (at least the versions of those pitches thrown by righties), and McCann is better at hitting curves than he is at hitting sliders both in terms of (i) quality of contact and (ii) not-whiffing. In other words, there isn’t a clear advantage to throwing six curveballs and zero sliders to this group unless you think your curveball is better than your slider. But if Mikolas thought his curveball was better than his slider, why not throw it more in his two starts?


To use a cliché, baseball is a game of adjustments. Each plate appearance is an exercise in game theory: the best strategy for the hitter depends on what the pitcher is going to do, and the best strategy for the pitcher depends on what the hitter is going to do. Did Mikolas thought that the Braves hitters would be sitting slider, so he gave them spinners that curve instead of slide? Maybe his arm hurt and he couldn’t throw a hard breaking ball. Or did he want to steal a few called strikes and thought his curveball would be the best to get them? I don’t know the reason.


What I do know: Miles Mikolas turned his season around when he began relying on his curveball more often in the second half. Part of being comfortable throwing four different pitches is that you can be unpredictable. Being unpredictable worked for Mikolas and the Cardinals when it mattered most.


Credit to @cardinalsgifs for the cover photo, and to Fangraphs and Baseball Savant for the various data sets referenced throughout this post.