Speed of sound: Statcast measures fastest outfielders

Who's the fastest outfielder in baseball? A few names immediately come to mind, like Billy Hamilton and Byron Buxton. Perhaps you're thinking about stolen bases, or highlight plays, or any of the ways in which we've traditionally defined "speed" over the decades.
But how do we know who is the fastest? Who's second, or third or 10th? Do outfielders run faster when they don't actually catch the ball? And how do you actually measure speed, anyway?
It's a question we've been thinking about a lot around the Statcast™ laboratory, and our answer is a metric we're calling "Sprint Speed," created by MLB.com's Tom Tango. It's got a very simple definition: How many feet per second did a player run in his fastest one-second window? On an individual outfield play, it's the basis of our speed-based range metric, Catch Probability. On a cumulative basis, it can create extremely interesting speed-based leaderboards. While we plan to apply this across a variety of running and speed-based situations, we're starting today by looking at speed during outfield defense because it was the natural next step after introducing Catch Probability.
The immediate question here is likely, "Why not just use miles per hour?" to express speed, because that's a scale most people are comfortable with. The truth is, we did that in the first two years of Statcast™, and it wasn't as illustrative as we hoped. Each individual baseball play takes only a few seconds, which means that expressing anything in terms of "an hour" is problematic.
Since expressing pitch velocity in mph has been the norm for decades, that's never going to change, but the convention isn't really helpful in other places. You'd never expect to see that a "500-foot homer" was actually a "0.094-mile homer," for example. You'd never say that the distance between bases was "0.017 miles," rather than 90 feet. It's the same idea with Olympic sprinters and prospective NFL draft picks, who are measured in terms of seconds over a set amount of feet or meters, not miles per hour. Everything else in baseball is measured in feet, so this ought to be as well, as this image makes clear:

In addition, when we did try to look at mph in the past, the results were unsatisfying because anyone could reach their "max speed" for a mere fraction of a second. Hamilton, for example, never showed up as being particularly impressive, because his skill is sustained speed, moreso than split-second reaction time. By looking at speed over the fastest one-second window, we're capturing approximately seven full-effort strides, and that's a much truer view of a runner's ability.
While we obviously didn't set out to create a metric to specifically showcase Hamilton, we knew that anything we did that accounted for speed would absolutely have to show him as being elite, or else we'd have known that there was something very wrong with our methods. As it turns out, the leaderboard we output is extremely satisfying. To generate this list, we looked at the 118 outfielders who had at least 100 fielding opportunities in 2016, found the top 5 percent of their speed times (to eliminate the many easy jogs each outfielder has, as Tango explained) on outs, and averaged that number.
Top 2016 Sprint Speed by outfielders
29.8 feet per second -- Hamilton
29.7 feet per second -- Buxton
29.1 feet per second -- Lorenzo Cain
29.0 feet per second -- Paulo Orlando
28.7 feet per second -- Jake Marisnick
28.6 feet per second -- Gregory Polanco / Travis Jankowski / Max Kepler
28.3 feet per second -- Christian Yelich
28.2 feet per second -- Kevin Kiermaier
Now we're getting somewhere, because these are very much the names you'd expect to see, and the next group of 10 includes more speedsters like Jarrod Dyson, Mallex Smith and Rajai Davis. At the low end, we find exactly the collection of sluggers, aging stars and injury concerns we'd expect -- Carlos Beltrán (25 feet per second), Matt Joyce (24.8 feet per second) and José Bautista (23.2 feet per second) pull up the rear, well below the 27 feet/second average of our 118 outfielders.
If these differences don't sound large, realize that on a fly ball with a hang time of five seconds, the difference between Hamilton going 29.8 feet per second and Bautista going 23.2 feet per second would be 34 feet, if they were going max speed for the entire time. They're unlikely to actually be doing so, but the point is to illustrate the scale, because a few feet per second matters.
For example, when Buxton robbed Melky Cabrera last year, he didn't just cover 90 feet in 4.7 seconds to make a Five-Star play with a mere seven percent Catch Probability, he also posted an elite sprint speed of 31.1 feet per second to do it. 

As you can see, for our qualified outfielders last year, the distribution clearly shows that the most-often sprint speed displayed was in the range of 27 or 28 feet per second, with the elite guys like Hamilton and Buxton far to the right.

(Where's Trea Turner, you're asking? He didn't play enough to qualify for the list, but if he had, his 28.8 feet per second would have placed him sixth, just above Marisnick, though he still wouldn't have been 2016's fastest rookie outfielder. That would have been Andrew Toles of the Dodgers, who also didn't qualify but would have placed third with a scorching 29.1 feet per second, as good as Cain.)
By placing everyone on the same "fastest one-second" scale, we're able to get past the obvious differences in distance that outfielders face on any given play, which isn't a concern in our examples of the Olympics and the NFL combine. But there's also another big difference in baseball, which is that it's not just about running as fast as you can. It's also about tracking a small white object that's moving at very high speed. So can we show if there's a difference between sprint speed on balls that were caught as compared to those that were not?
We sure can. Looking at our group of qualified players, we can see that they averaged 27.0 feet per second when making the catch, and 27.7 feet per second when going after balls they didn't get to. Intuitively, that makes sense, because that one-second window may fall while a player is hustling as fast as they possibly can to get a ball that's already past them, without worrying about tracking where the ball may land. Very few of our outfielders were faster on average when making the catch than when they did not; only nine had differences of at least 0.3 feet per second. That's why we're using "on outs" as our standard, as it's a better measure of 'baseball speed,' and isn't that the point?
Since we have two years of Statcast™ data, we can also compare outfielders who played in both seasons and see how well this metric correlates from year-to-year. As it turns out, it works pretty well, which makes sense -- barring injury, it's not like Hamilton or Buxton are going to suddenly stop being fast, right? There were 79 outfielders who had at least 100 opportunities in both years, and as you'd expect, the year-to-year relationship is strong.

The outlier at the left is Bautista, and that makes a great deal of sense. At 35 years old last year, he missed considerable time with injuries to both his left toe and his left knee, making a decline in speed something like the least surprising outcome possible. He lost 1.8 feet per second from 2015 to '16, dropping from 25.1 feet per second to 23.2 feet per second, and we're fine with someone having a different performance year-to-year if there's obvious health-related reasons for it.
You can expect to see more and more of this as the year goes on, but what's next? Well, so far, we only looked look at this through the prism of outfield defense. The next step will be to do the same for baserunning, and that opens up exciting opportunities. How will infielders like Dee Gordon and Jonathan Villar measure up? Does Hamilton run faster on the bases or in the field? Which speedy runner isn't great at stealing bases? And are there slower players who are excellent at it? Now, we have a better tool with which to answer those questions.