Byron Buxton, Tim Anderson, and the Judicious Application of Speed by Ben Clemens July 15, 2022 Kelley L Cox-USA TODAY Sports Baseball is a dynamic sport. Sure, it’s static plenty of the time — there’s a lot of spitting, scratching, and adjusting between pitches — but wait for a ball in play, and everyone on the field moves as though jolted by electricity. Fielders rearrange themselves based on the path of the ball. Baserunners charge, pause, and retreat according to a set of rules they know by heart. The batter tears out of the box — or jogs, or strolls, as the case may be — before adjusting based on the flight of the ball and the action on the field. That intricate burst of motion is thrilling. It explains why fans, players, and the league as a whole want more balls in play. Thanks to Statcast, it’s also measurable. More specifically, Statcast measures and publishes sprint speed, which you can think of as the average top speed a player achieves across all of their competitive runs. Because it’s directly measuring what’s happening on the field, the sprint speed leaderboard mostly matches what our eyes see. The fastest player in baseball? That’d be Bobby Witt Jr., barely edging out Trea Turner for the top spot on the list. Muscle-and-hustle standouts Oneil Cruz and Tyler O’Neill are near the top. Three catchers (Yadier Molina, Martín Maldonado, and Yasmani Grandal) are tied for last. The board works! On the other hand, maybe the board doesn’t work. An average of your competitive runs depends on what “competitive” means. Loaf on one of those, and you’re dragging your average down. Loaf on a run that just ticks over the border to uncompetitive, and it won’t count. Anytime you’re measuring a subset of all runs — an admirable goal, because comparing apples to apples is important — you’re opening yourself up to error due to the definition of that subset. Luckily, Statcast also measures “bolts,” or runs where a player’s top speed exceeds 30 feet per second. As you might imagine, faster players eclipse that threshold more often. Witt, for example, has topped 30 ft/sec on 55% of his competitive runs. Trevor Story, who is slower but still fast (average sprint speed of 28.4 ft/sec), has topped the 30 ft/sec threshold on 3.2% of his sprints. Jazz Chisholm Jr. is roughly halfway between Witt and Story in average sprint speed, and he’s also between them in bolt rate; 17.1% of his competitive runs have been bolts. Among players who average 27.4 ft/sec or better for sprint speed — I cut the data off there because there aren’t really any bolts below that — bolt rate and sprint speed have a roughly quadratic distribution: The faster you are, the more frequently you “bolt,” obviously enough. That surely tops out somewhere; you can’t bolt more than 100% of the time, if nothing else. But no one currently in baseball is fast enough to test that threshold. The correlation between bolt rate and sprint speed is tight — above a 0.8 r-squared — but it’s not absolute. Again, this sounds like real life. Imagine you’re on a recreational softball team, and your two fastest players behave differently. One, a real try-hard, goes all out on every run, chains flying and helmet popping off as he busts it down the line on every grounder. Another player is faster, but she’s also lazier; she only puts the pedal to the metal if she expects a bang-bang play at the base. The first player could have a higher sprint speed, and the second player has a higher bolt rate (or whatever the equivalent version of that in slow-pitch softball may be). At the major league level, I don’t think anyone is truly lazy. You don’t get that far in a wildly competitive field by loafing. But players exert different effort levels from play to play for various reasons: conserving energy, preventing injury, or simply understanding that an all-out sprint isn’t called for. Let’s take Jose Altuve as an example. Until recently, his speed could best be clocked in that Chisholm range, between 28 and 29 feet per second on average. He routinely posted bolt rates that were higher than that, though; for the first six years of the Statcast era, as his speed waxed and waned, he kept posting top-end bolt rates: Jose Altuve, Sprint Speed and Bolts by Year Year Sprint Speed Bolt Rate 2015 28.5 22.5% 2016 28.4 17.8% 2017 28.4 19.0% 2018 28.1 20.2% 2019 28.6 18.3% 2020 28.4 18.9% 2021 28 16.7% 2022 27.4 9.1% Even this year, with his speed cratering, Altuve has something in the tank when he needs it. He’s barely on the graph up above, hanging on at 27.4 ft/sec, but his 9.1% bolt rate is the best mark among players below 28 ft/sec. He’s just plain faster, at his true top speed, than players with the same average sprint speed. He merely breaks it out less frequently; he’s getting older, after all, and part of the wisdom of age is knowing when to punch it and when to conserve energy. Altuve’s a wily old veteran (somehow only 32, but that’s old for fast players), but younger players do the same thing. Take Byron Buxton, one of the best athletes in the game. Buxton has slowed down significantly this year, with his average sprint speed falling by 1.1 ft/sec: Byron Buxton, Sprint Speed by Year Year Sprint Speed 2015 30.9 2016 30.8 2017 30.6 2018 30.5 2019 30.4 2020 30 2021 30 2022 28.9 In fact, players with a sprint speed near Buxton’s yearly average of 28.9 ft/sec record very few bolts overall. On average, they’ve exceeded 30 ft/sec on 10% of their runs. That makes sense based on the numbers we heard earlier; 10% is roughly halfway between Chisholm and Story, for example. Buxton doesn’t fit that mold: Byron Buxton, Sprint Speed and Bolts by Year Year Sprint Speed Bolt Rate 2015 30.9 72.1% 2016 30.8 82.8% 2017 30.6 87.6% 2018 30.5 58.8% 2019 30.4 46.9% 2020 30 50.0% 2021 30 36.4% 2022 28.9 26.1% What does it mean? On average, Buxton has less pace than Adam Engel, Owen Miller, and Brandon Marsh. When he needs it, though, he’s faster, because of course he’s faster. He’s Byron Buxton! At maximum effort and at full health, he’d dust those three in a race. That shows up in his bolt rate even as his average sprint speed has declined, perhaps as a result of a conscious effort to pace himself and stay healthy. The best example of this phenomenon, though, is Tim Anderson. Anderson isn’t especially fast. He’s never averaged 29 ft/sec over a full season, and his sprint speed mark has been lower the past two years: 28 ft/sec in 2021, 28.1 so far this year. But take a look at Anderson and the other players with the same average top speed: Sprint Speed and Bolts, Tim Anderson and Comps Player Sprint Speed Bolt Rate Steven Kwan 28.2 6.5% Shohei Ohtani 28.2 3.6% Alcides Escobar 28.2 2.3% Tommy Pham 28.2 1.0% Harold Ramírez 28.2 0.9% Christian Arroyo 28.2 0.0% Ernie Clement 28.2 0.0% Willi Castro 28.2 0.0% Tim Anderson 28.1 19.7% Austin Slater 28.1 7.3% Nick Gordon 28.1 1.5% Luis Rengifo 28.1 1.3% TJ Friedl 28.1 0.0% Luis González 28.1 0.0% Eric Haase 28.1 0.0% Avisaíl García 28.1 0.0% Dylan Carlson 28.1 0.0% Cody Bellinger 28.1 0.0% Matt Chapman 28.1 0.0% Tyler Naquin 28.1 0.0% He’s just faster, no matter what sprint speed tells you. In fact, if you’re thinking about pure speed, bolt rate might be more useful. It suffers from an arbitrary threshold; if you can’t post a 30 ft/sec time, you’ll show up with a 0 bolt rate whether you’re a Molina brother or a garden variety slow boy. But how frequently you can turn in a fast run is a great indication of how fast you are, more so than how often you put in maximum effort. Other takeaways from this? I’m not quite sure how bolts interact with home-to-first time; someone with more data sense than me should absolutely look at that. I’m not sure what the competitive run threshold is, though it seems to work just fine for establishing who’s fast and who isn’t. There’s a lot more to be learned about how fast and how often players sprint. But I find the bolt data interesting, and I’ll watch Anderson, Buxton, and Altuve with new appreciation. Those guys are fast — and the data might still underestimate their speed.