Blake Snell and the Terrible, Horrible, No Good, Very Bad Luck by Ben Clemens July 1, 2019 On November 14, 2018, Blake Snell won the AL Cy Young award. It was a close vote, but no one could say Snell didn’t deserve at least to be in the discussion. He compiled a 1.89 ERA, best in the AL, and his peripherals (2.95 FIP, an outrageous 31.6% strikeout rate) weren’t far behind. He was, simply put, one of the best starters in baseball — unfair, as future Rays employee Jeff Sullivan put it. Just more than seven months later, on June 29, 2019, Blake Snell’s ERA was on the wrong side of 5. By RA9-based WAR, he was barely above replacement level in 2019. A strong start yesterday moved his ERA down to a still-inflated 4.87, but it’s worth asking: is something wrong with Blake Snell? Now, as my RotoGraphs colleague Al Melchior recently put it: nothing is wrong with Blake Snell. Still, it seems like it might merit investigating. Guys with stuff like Snell’s aren’t supposed to even be capable of putting up near-5 ERA’s this far into the season. Al focused on Snell’s strike-throwing, and that’s always a make-or-break issue for a guy with such dynamite stuff, but Snell’s walk numbers, while high, aren’t crippling. He’s actually walking fewer batters than last year, and his K-BB% is a career high. No, Snell’s 2019 has been alarming because of his inconsistency, and that’s worth looking into. In 2018, Snell made only four starts in which he didn’t last at least five innings. One was his first start back from injury, which hardly counts. This year has been an entirely different story. Snell’s start on June 25, when he survived only 3.1 innings against the Twins, was his sixth outing of 2019 to see him not finish the fifth inning. There’s always batted-ball luck involved in short outings, but still, Snell’s 2019 feels extreme. Did he change something in 2019 that’s leading to more abbreviated outings? It’s worth saying again that Blake Snell is incredible. All four of his pitches are weapons. His four-seam fastball is the fastest thrown by any left-handed starter, and it generates whiffs on more than a quarter of batters’ swings against it. Its rise and fade are near-unmatched; only Justin Verlander gets more total movement on his four-seam. Snell’s curveball, which he’s throwing 27% of the time this year, is awe-inspiring. Batters whiff on 55% of their swings against it, the second-best mark for any starter who has thrown 100 curveballs this year. His changeup? It generates the fourth-most whiffs per swing, 44%. He rarely throws his slider (7.6% of the time so far this year), but you guessed it: no starter’s slider gets more whiffs per swing than Snell’s. What do you get when all three of your secondary pitches are among the best in baseball, and they back up a Verlander-esque rising fastball? You get swings and misses like you wouldn’t believe. Take a look at the starters with the best swinging strike rates in 2019: Best Starter SwStr%, 2019 Pitcher SwStr% K% ERA FIP Blake Snell 18.3 33.3 4.87 3.34 Max Scherzer 16.7 34.8 2.43 2.09 Justin Verlander 15.8 32.1 2.86 3.99 Gerrit Cole 15.8 36.8 3.28 3.10 Luis Castillo 14.9 28.5 2.47 3.79 Numbers two through five on this list have combined for an ERA of 2.78. Snell is comfortably ahead of them when it comes to getting swinging strikes. The point is, Snell is working with unmatched stuff. Some of Snell’s ERA comes from the usual bad-luck problems. His left-on-base rate (the percentage of runners left stranded) is the seventh-lowest among qualified starters. Still, that can’t explain the FIP-ERA gap: the six pitchers with a lower LOB% have, on average, ERAs 0.61 higher than their FIPs. Snell’s ERA-FIP gap is a whopping 1.53. Clearly, there’s more to the story. Another classic driver of unlucky pitching performances is BABIP. The key insight behind FIP is that batted balls become hits more or less at random, but that’s been far from the case so far this year. Snell’s BABIP of .354 is the highest in baseball by 15 points. If his BABIP holds for the rest of the year, that would make him historically unlucky: only one qualified starter since 2000 has ever posted a higher BABIP (2008 Kevin Millwood). With the advent of Statcast, however, we can do better than just say that Snell is probably getting unlucky. We have the exit velocities and launch angles for almost every batted ball in baseball. Batting average might not distinguish between a smashed line drive and a seeing-eye single, but Statcast’s expected stats do. Are batters simply squaring Snell up so frequently that his high BABIP is merited, à la Jose Fernandez in 2016? Think again. Snell’s xBABIP, the expected percentage of non-home run batted balls that land for hits, is .311. He’s worse than average in that department, but only barely; he’s in the 45th percentile among qualified starters. Quality-of-contact metrics agree. Snell allows a below-average 87.4 mph exit velocity on average. Batters are hardly hitting balls in the air with authority; his 89.3 mph exit velocity on fly balls is one of the lowest marks in baseball. It’s not a matter of a bunch of crushed line drives and a few pop ups, either: batters have barreled up 5.4% of the balls they’ve put in play against Snell, a lower rate than 90% of pitchers who have allowed 100 balls in play. Like your metrics old-school? Snell’s hard-hit rate of 33.8% is excellent, well below the league average of 37.9%. Taking that fact and the barrels into account, he can probably expect his home runs per fly ball to decline; his 18.5% rate is 10th-worst among qualified starters, and nearly every pitcher in front of him has a worse hard-hit rate (Yu Darvish, your 2019 is very weird!). He’s getting more groundballs, too; his GB/FB ratio of 1.32 is the highest of his major league career. Look at these batted ball metrics, and Snell looks better than last year, not worse: Blake Snell’s Contact Quality Year EV EV (air) Barrel/BBE Hard Hit % GB/FB 2018 87.0 90.0 7.2% 35.7% 1.23 2019 87.4 89.3 5.4% 33.8% 1.32 There’s almost no way you could look over Snell’s batted ball metrics and expect anything worse than near-average contact management. Yet here he is with a historically high BABIP, watching 29% of the groundballs hit against him find holes (league average is .240) despite completely unremarkable exit velocity. Is Snell doing something to allow hits that none of these metrics pick up? Maybe! It’s quite unlikely, though. Splitting 2018 into two halves, there was no correlation (r-squared less than .01) between first-half and second-half BABIP-xBABIP. In other words, being unlucky relative to “expected” really does seem like luck, rather than some effect the formula is missing. Is this all we need to say? Blake Snell is still great, nearly as great as he was last year, and he’s just getting unlucky? It’s tempting to leave it at that. He has tremendous fielding-independent numbers, and it appears that he’s doing everything in his control to keep contact against him from being valuable. That doesn’t quite tell the whole story in my mind, though. There’s still the matter of the short starts. Can those be explained away by the same luck? Can we expect Snell’s consistency to come back when a few more balls land in gloves? Of Snell’s six short starts this year, he’s had a Zone% lower than his season average in four, according to Baseball Savant. Three of those were his lowest three zone rates of the year, in fact. In those three starts, he had a 19.1% walk rate. Is he just wild more often this year than last year, missing his spots so much that batters either walk or bat in favorable counts constantly? It seems not. Snell’s three lowest-zone-rate starts of last year had even lower zone rates than 2019. In those starts, he completed 18.2 innings with an 11% walk rate and a 39% strikeout rate. Clearly, Snell doesn’t have to be in the zone to succeed. Is it a pitch issue, perhaps with one pitch losing effectiveness and problems cascading from there? This seems a little more believable. Snell’s curveball is spinning less than it ever has — his six June starts comprise six of the seven lowest spin rates he’s had on his curve ball this year, and four of those are lower than any start from 2018. Snell seems to trust the pitch less — he’s thrown it less often than in any other start this year in his last two starts, leaning instead on his fastball. That plan has had mixed results — a career-high zone rate, 11 hits, 10 swinging strikes, and only four strikeouts one start; three hits, 20 swinging strikes, and 12 strikeouts the next. I always try to be careful when doing such granular pitch-level analysis. It’s easy to delve too greedily and too deep, to lose yourself in specifics and see a pattern behind every fastball. What’s most likely, I think, is a far more mundane answer. Snell doesn’t quite have the feel for his curveball that he’d like, and so he’s been throwing it less. He’s experimented with what to replace it with, getting mixed and inconsistent results. His other pitches are spinning a bit less as well, and his fastball velocity is marginally down; perhaps he’s just fatigued. In any case, any effects of the change in pitch mix have been swamped by luck. No one can be an effective pitcher when they allow the BABIP that Snell has. He could be exactly the same pitcher as last year and we’d never know, so different have his results been after contact. What will Blake Snell look like in the second half? I’d keep one eye on his curveball usage, but until something else changes, my expectation is the same as it was before the year: Snell will be one of the best pitchers in baseball, with otherworldly stuff, occasional control problems, and a bright future. One half-season of the worst luck imaginable isn’t enough for me to change that view.