Braves hope Drew Smyly can ace his 2021 test

When the Braves signed Drew Smyly on November 16, 2020, it came as a bit of a small shock. It was the second time in two years that General Manager Alex Anthopoulos pulled the trigger on scooping up an erstwhile San Francisco Giants hurler shortly after the conclusion of the World Series... but unlike his previous adventure in this regard (Will Smith, $40 million guaranteed over three years), there was no Qualifying Offer timetable forcing anyone’s hand. That was actually the first free agent signing of a major leaguer to a major league deal of the offseason (the Giants signed Jason Vosler, who has yet to make his MLB debut, to a major league deal six days prior).

The sum was perhaps a little curious: $11 million. At the time, there was a morass of uncertainty regarding team budgets and spending in the wake of the COVID-19 pandemic, but $11 million outdid most estimates. The Fangraphs crowdsource saw Smyly getting a one-year deal at $5 million. MLB Trade Rumors’ prediction was the same. Craig Edwards at Fangraphs saw Smyly locking down a two-year deal, but only at $8.5 million per season. That the Braves were the ones to give Smyly a higher-than-expected deal, in this economy, came as a bit of a stunner. Further market developments didn’t help in the clarity sense, either. In the tier of starting pitchers projected to perform basically as well as Smyly, the assortment of Adam Wainwright, J.A. Happ, Garrett Richards, Martin Perez, Alex Wood, Jose Quintana, Anthony DeSclafani, Robbie Ray, and Chris Archer all got less than Smyly’s $11 million. In fact, at the time of writing, the only free agent pitchers to beat Smyly’s guarantee are Charlie Morton (also signed by the Braves), Mike Minor (a two-year deal with a $9 million average annual value, or AAV), and of course, Trevor Bauer. (Kevin Gausman and Marcus Stroman both accepted qualifying offers.) If you had Drew Smyly getting the third-biggest starting pitcher AAV of this free agent class, raise your hand, you genius you.

In any case, what’s done is done, and the Braves are looking forward to Smyly getting a substantial number of outs for them this coming season, presumably while also pitching well in the process. The main question in this regard is “how,” and the answers are anything but straightforward. To cash in on his upside, the Braves and Smyly will need to get answers to a bunch of queries, some of which we’ll focus on for the body of this post.

---

Can Smyly stay on the field?

In some ways, Smyly’s fantastic, small-sample 2020 was not a surprise given his track record. A second-round pick back in 2010, Smyly rocketed through the minors and won a rotation spot in Spring Training. His performance was real good for a rookie: 106 ERA-/93 FIP-/94 xFIP- in his first 15 starts. He finished the season at 96/92/96 in 18 starts and five relief appearances in 99 innings, racking up a really impressive 1.8 fWAR in the process. The next season, he was a multi-inning relief weapon, again finishing with 1.8 fWAR, which was a top-10 reliever season. Over the next two seasons, he continued to rack up solid stats. In short, through his first four campaigns, Smyly had nearly 7 fWAR in under 400 innings, with an enviable 81/90/92 pitching triple-slash. I highlight this to emphasize that the Drew Smyly story isn’t one where a hangaround journeyman has become a desirable asset because of a great-but-short 2020 campaign. Rather, 2020 was a (brief) return to form (and then some) for the lefty, not the first time he showed he was capable of mowing down batters.

The problem is, of course, that things really fell apart for Smyly, starting in 2016. That was his first below-average season, nearly across the board, as he went from “pretty damn good” to “worse than mediocre” in everything but xwOBA-against. Notably, he didn’t lose velocity, or change his pitch mix, or anything else. Heat maps show that the command of his breaking pitch and cutter got considerably worse, and that hitters elevating and victimizing the latter may have been an issue... but whether he could address this issue in the future quickly became a moot point, as Smyly ended up missing the next two seasons with the whole elbow-injury-becomes-Tommy-John-Surgery thing. By the time he got back on the field in 2019, he was still throwing at his usual 91ish, but everything else was mostly a mess as his fastball and cutter were getting obliterated. After being released by the Rangers, Smyly cut down on his fastball usage and had a much more successful 12-start stint with the Phillies, though he still struggled to hit the zone consistently. His 0.6 fWAR (with a 96 xFIP-) in those 12 starts couldn’t salvage his stint with the Rangers, and he finished at -0.3 fWAR for the year.

Even his pseudo-renaissance with the Giants was not free of health-related concerns. While Smyly was fantastic when healthy, he missed over a month of what was a two-month season with a finger issue, and only averaged around 13 outs per start. This is therefore, plain and simply, likely the biggest concern with Smyly: can he hang around to provide value over the course of the season?

Projections generally struggle with playing time and health considerations, at least relative to forecasts of player performance. Steamer’s central estimate is 24 starts totaling 134 innings. ZiPS has 98 innings across 18 starts and three relief appearances. I don’t envy anything having to prognosticate like this — I have no idea what the “right” way to forecast playing time for a guy who missed two of the last four years in their entirety, missed a bunch of time in the most recent season, and was averaging only 114 innings a season for various reasons even before 2020. It would not surprise me if Smyly was really good, inning-for-inning and pitch-for-pitch, in 2021. He’s done it before, even if a lot of “before” was a long time ago. It would surprise me if he’s able to stick around for substantially more than 90-100 innings in 2021, but given that the Braves committed $11 million to him, that seems to be what they’re banking on.

Bottom line: is 2020 Smyly for real?

Something that might make the “hey, this guy might be on the shelf for a while” pill easier to swallow is if Smyly continues to overpower hitters when he’s able to take the mound. Assuming 100 innings at his career fWAR rate would give you a 1.2 fWAR pitcher. Steamer and ZiPS are a little less optimistic: 0.9 to 1.1 per 100 innings. Yet, in 2020, Smyly racked up 1.0 fWAR in just 26 (!!!!!!) innings. Only five players had a better FIP- than Smyly’s 48 in 26 or more innings, including just two starters. Only three had a better xFIP- than Smyly’s 58 with more innings, and two of those were Shane Bieber and Jacob deGrom.

It’s difficult, if almost impossible, to look at Smyly’s seven-game sample and feel confident about just carrying it forward into the future. But, this isn’t to say that Smyly was just exceedingly fortunate to have a good seven-game stretch that happened to comprise the entirety of a shortened season. There’s two cases to be made here: the numerical case, and the mechanical/pitch quality one.

From the numerical side, we can see that in Smyly’s career, seven-game stretches like the one he had in 2020 have been fairly rare — even though he was quite a good hurler. As far as xFIP- goes, there were a few dominant relief appearance stretches in 2013, but that’s really about it. The two red lines below cabin his rolling xFIP between the start of 2020 and the end — you can see that aside from his full-relief season, he hasn’t generally even been in the ballpark of where he was after a few 2020 outings, much less after the full seven-count suite of them.

I won’t post a bunch of graphs, but a seven-outing stretch this good isn’t common in general. Jacob deGrom’s only had three (one was 2020). Shane Bieber’s career is very young, but he’s only managed it for about a month in 2019 and then for about six weeks in 2020. I’m not saying Drew Smyly is anywhere near those guys — just that his 2020 was very good in its teeny sample, and muddies the water a bit in terms of what to expect. (Meanwhile, Clayton Kershaw is a monster and spent like three whole seasons in this area, so good luck 2014-2017 batters.)

But, this analysis is predicated on Smyly having a constant, central talent level, with deviations in small samples in one direction or another. It could very well be the case, though, that 2020 should suggest that his talent level has actually changed. In other words, what if what’s past is past, and doesn’t matter for future-Smyly? Consider:

• Smyly’s fastball averaged 93.8 mph in 2020. That’s nearly 3 mph higher than 2019, more than 3 mph higher than the three years before his injury woes, 2 mph higher than his full-relief season in 2013, and nearly 2 mph higher than his rookie season. Only Felix Peña gained more on his fastball between 2019 and 2020 (and no one else was particularly close), and Peña didn’t start a single game in 2020. In short: Drew Smyly is the only “primarily a starter in the second year” guy with a 2.5 mph or greater gain on any fastball (which includes his cutter) going back to 2015-2016. All the others are either guys who substantially increased their relief appearances in the second year, or just pure relievers like Scott Barlow. The point is that Smyly really gained some extra velocity, to an unusual extent.
• We’ve already talked at length about Smyly’s bizarro curve-that-works-like-a-screwball. The Braves are aware too, to the extent that Alex Anthopoulos directly mentioned it on the radio after the signing. Up through 2019, Smyly only threw it around a quarter of the time, even though his xwOBA allowed on the pitch (since 2015; no xwOBA available before then) was .240, while his xwOBA allowed on everything else was .359. (For comparison, the league tends to hit .260 on breaking pitches and around .340 on non-breaking pitches, so Smyly was arguably under-using a better-than-average breaker despite worse-than-average other stuff. Smyly threw 1.4 curves in 2020 for each curve he threw (proportionally) earlier in his career, and around 1.2 for each curve he threw in 2019. The point is, if the curve is really the weapon it appears to be, and the increased usage is a new normal for Smyly, perhaps the past doesn’t matter as much. Note that his 2020 usage was around 46/36/18 fastball/curve/cutter — plenty of room to increase curveball usage further.

There are some corners of baseball-analysis-land that will point to other factors here, regarding a changed delivery. I can’t really confirm or deny that these matter, both because I don’t know what to look for, and also because the numbers in this regard are ambiguous or worse. It is true that Smyly got more extension in his delivery, which made his effective velocity play up. But, I’m not sure this matters too much. For one, Smyly gets below-average extension in his delivery, and his “effective velocity” underperforms his out-of-the-hand velocity relative to a league-average delivery. He did improve it in 2020, going from a 1.4 mph deficit to an 0.6 mph deficit. But while his relative gain in this regard was one of the biggest in baseball from 2019 to 2020, there’s essentially no relationship between “improvements to extension” and “improvements in xwOBA-against.” As such, it’s hard to say this is part of any mechanical case. Still, maybe you can see some useful differences in the below — I can’t! (As a side note, the Braves’ other rotation addition, Charlie Morton made huge changes to his delivery extension in 2020, going from below average to slightly above average — and fared much worse with his fastball!)

2019:

2020:

Both pitches are down-the-middle fastballs to lefties at the same venue for comparison purposes.

These two things (health, effectiveness level) are the obvious high-level questions for Smyly, but that’s hardly surprising: they’re the same questions you’d ask about pretty much any player at the start of any season. It’s just that given that Smyly missed two seasons recently, was terrible in another, and then fantastic for 26 innings in a fourth makes this all very complicated.

With that said, there are a few more questions worth asking here. The answers to these may help clarify whether Smyly will give the Braves what they wanted in 2021.

The curveball will still be elite, right?

The Smyly addition was likely all about the curveball. The stats are ridiculous in the small 2020 sample — starting and ending with the fact that hitters whiffed at it literally half the time they offered. Among all pitches that ended 50 or more PAs in 2020, Smyly’s curve was top 10 in whiff rate, strikeout rate, and putaway rate — while being a top 25 pitch overall in xwOBA-against. The one place where it didn’t succeed was that hitters tended to sock it fairly well when they did connect, 92.1 mph of exit velocity, but when they’re not even connecting half the time, who cares?

Smyly’s curveball has generally beffudled hitters — the highest single-season wOBA he’s allowed on it, ever, was the .294 mark in 2019. For comparison, the league’s wOBA on curves is around .260, something Smyly’s curve has generally outperformed, aside from his pre-injury 2016 and post-injury 2019 (and also 2015). Going back to 2015, we also have xwOBA — the league’s xwOBA has been consistent with that .260 mark, while Smyly’s put up 2015/2016/2019/2020 marks of .243/.215/.263/.205. Basically, at his worst, the curve was still an average curve in xwOBA terms. At best, it’s been dominant.

Seen in the context of earlier seasons, his 2020 seems better, but not necessarily dramatically different. He posted a 40 percent whiff rate on the breaker (classified by Statcast as a slider, but validated as a curve) in 2016. You might think that the difference could be location-related, i.e., Smyly’s curve jumped in effectiveness because he was better at spotting it where hitters would offer but miss. That’s not really the case: Smyly actually threw his curve in the heart of the zone at a career-high rate. Instead, the difference is that when he did throw it outside the zone, it was essentially unhittable: whiff rates on the pitch were 51 percent or higher on the edges or further out. Yet, again, the overall ratio of curves thrown out of the zone was fairly low. It’s not clear whether this is inherently replicable — it’d be an easier story to tell if it involved higher whiff rates everywhere, or more curves thrown where whiff rates are higher. But with neither of those being the case, it’s an open question of whether this will keep up, or is just a small sample artifact. I guess we’ll see.

There’s also that whole pesky regression-to-the-mean thing, as it applies to basically everything. Of the 17 pitches with whiff rates of 50 percent or higher (minimum 50 PAs) from 2019 that also appeared in 2020, only six had whiff rate increases the following season, and the average whiff rate change was a decrease of 5.6 percent. More of these pitches had double-digit whiff rate decreases than increased in any way. (Side note: good lord, Tanner Rainey’s slider, what a pitch.) If you look at it in terms of xwOBA, pitches in the 95th percentile of eliciting the best (worst for hitters) outcomes in 2019 had xwOBAs ranging from .121 (Tyler Glasnow’s curveball) to .205 (I set the cutoff here because this was Smyly’s curveball xwOBA in 2020). The average xwOBA-against for these 56 pitches was .185. In 2020, these same 56 pitches allowed a .234 xwOBA — more than twice as many got worse xwOBA-wise than got better, with an average change of .049. If that happens to Smyly’s curveball in 2020, it’ll go from devastating to a league-average breaker, still effective and part of a decent hurler’s arsenal, but not much more.

So, as with his overall performance level, Smyly and the Braves face more or less the same question with his curve: can you really buy 2020, or are you prepared for something that looks like his entire career? The curve going from best-in-class to average won’t break Smyly in any way — but it will make the signing look more curious in retrospect. At least, no matter what happens, it’ll probably still be backwards, and that’s amazing in and of itself.

All this aside, though, Smyly has a clear path to betterment by just throwing his curve more, even if the curve doesn’t carry over all of its 2020 qualities and outcomes. That alone should budge up his performance relative to 2019 and his career, and seems eminently doable.

Is there anything more to unlock in the four-seamer?

On paper, Smyly’s four-seamer, especially the 2020 version, seems awesome.

Smyly’s fastball was both harder than average and had more “rise” than average. Even after accounting for the fact that harder fastballs have less time to drop, Smyly was just about in the top third of “rise” for pitchers with 220+ four-seamers thrown last season.

And yet, his results on the fastball were pretty uninsipring. Among players with 45 or more PAs ending on a four-seamer in 2020 (45 is Smyly’s count):

• Bottom third in xwOBA allowed (.361)
• Average in whiff rate
• 40th percentile in strikeout rate
• Nearly bottom fourth in put-away rate (rate of two-strike pitches resulting in a strikeout)
• Bottom third in hard-hit rate allowed

(Note that the percentiles above are not raw data but value judgments, i.e, bottom third is bad.)

There are a lot of hypothetical reasons for this, but it’s hard for any of them to bear out as particularly informative given the small-sample nature of 2020. You could point to Smyly’s location (it’s not very up), but guys with very similar fastballs in both velo and rise, and much better fastball xwOBAs, like Zac Gallen, don’t throw up either. And there’s Grant Dayton, who locates similarly to Smyly’s with a much slower fastball, yet manages to bamboozle hitters better with it anyway.

In short — Smyly’s fastball doesn’t quite seem to work as well as it should. It remains to be seen whether he can align its outcomes to its pitch quality, or whether it’ll matter less and less because it will take a backseat to his curve in 2021. The key thing here is that Smyly’s ,361 xwOBA-against on the fastball is below-average for fastballs (league average is around .340). If he can figure out a way to get more out of the heater, whether by throwing higher in the zone or something else, it should pay substantial dividends. Without it, he’ll have to rely more heavily on his curveball, and remain kind of a quirky guy in that his great-on-paper fastball somehow plays substantially down.

---

So, there you go. Four questions for Smyly that will govern his 2021 season:

1. Can he stay healthy?
2. Was 2020 a small-sample effect of his prior, average-y talent level, or did he find a new gear with the added velocity, spin, pitch mix, and other things? In other words, is he a different pitcher now, or did the 26 innings constituting a whole season for Smyly just make it look like he became a different pitcher?
3. What was behind his curve’s success in 2020 compared to prior years? Can he replicate it, or will regression to the mean make his best pitch simply a good one rather than an elite offering?
4. Can he unlock anything more from his four-seamer, a pitch that seems like it should be much better than it has been?

We’ll find out together — and hopefully the answer to the first question is “yes,” so that we can explore all of the others.