What Can Speed Do?

Over at Baseball Savant, another Statcast leaderboard has been rolled out. This one relates to speed. They are calling it Sprint Speed, and the definition is as follows:

Sprint Speed is Statcast’s foot speed metric, defined as “feet per second in a player’s fastest one-second window.” The Major League average on a “max effort” play is 27 ft/sec, and the max effort range is roughly from 23 ft/sec (poor) to 30 ft/sec (elite). A player must have at least 10 max effort runs to qualify for this leaderboard.

While Sprint Speed has been used for a while, we didn’t have leaderboards until now. Mike Petriello over at MLB.com has a full article on the rollout which I would recommend. Among the highlights: Sprint Speed correlates well from year to year; it doesn’t require a large sample to become reflective of true talent (Petriello compares speed to fastball velocity); and it might be useful when attempting to identify injuries that could be slowing players down.

So, we know that the metric can tell us who is fast and who is not. That’s helpful. I wondered if it might also be able tell us anything about any other statistics.

Before trying to predict the future or look at past years, I thought it might be useful to compare speed to the stats we have and see how they compare. While the leaderboard over at Statcast features nearly 350 names (every player who’s produced 10 or more max-speed data points), those sample sizes might be a bit too small when looking at other statistics. As a result, I narrowed the sample for this study down to the 166 players who were qualified at the end of Monday’s games.

I was concerned that cutting the number of players sampled in half might skew the data. Removing part-time players, I thought, might take away too many quick, defense-first players who didn’t play full-time — as well as quite a few catchers and lumbering platoon sluggers who might not qualify. I compared the overall averages of the 350-player sample to the qualified sample and the averages were nearly identical. I also looked at the top quarter of players in each sample and the bottom quarter of players in each sample, and those numbers were nearly the same, as well.

Before getting to some of the more speed-based stats, let’s take a quick look at overall value and see if there’s any relationship between speed and value, looking at this year’s WAR.

There’s not a strong relationship here, but this shouldn’t come as a surprise. Speed is just one tool and reveals little about whether a player can hit, hit for power, field, or throw — or any of those other tools which have considerable bearing on a player’s value. Before we get to steals and baserunning, there was a stat about which I was curious and that relates to defense. Are fast players better on defense? We would think so in the outfield, but would it make a difference all over the diamond?

We see a bit more of a relationship here than we did with WAR, but it isn’t incredibly strong. A few caveats within this group, however. First is the standard defensive caveat that UZR takes time to stabilize (although, including the positional component blunts that somewhat). Second, catchers are going to be good at defense and likely slow, which skews the data slightly. If you simply remove the seven qualified catchers from the group, the correlation coefficient jumps from .30 to .39. Likewise, if you just look at outfielders, the correlation coefficient jumps further, up to .45. This says a couple things: (a) there’s likely some relationship between defense and speed, in the outfield in particular and (b) there’s more to defense than speed. Positioning, jump, hands, and arm all play a role as well.

How about stolen bases?

We see a pretty good relationship here. The faster a player gets, generally the more stolen bases he’s recorded this season. The fastest player without a steal this season is Carlos Correa, while Albert Pujols is the slowest player in baseball who has a steal. (His former teammate and still very slow Yadier Molina has five.) There’s only one player in double-digits below 27 feet per second, Paul Goldschmidt, and he is close to the average of 26.9. Being fast doesn’t guarantee a lot of steals and being slow doesn’t mean you can’t steal, but speed definitely helps.

Speed helps you steal, but does it help you not get caught, too? Stolen-base runs (wSB) accounts for steals and caught stealing.

There’s a relationship, but it isn’t as strong as just looking at stolen bases. This is likely the result of some fast players believing they are faster than they are and some slow players realizing that they are slow. What about just speed on the basepaths without regard for steals? Ultimate Base Running (UBR) measures runs produced on the bases (but not steals).

The relationship here is stronger than that of wSB. When running on batted balls, speed is going to have to come in to play — unlike stolen-base chances, where slower players can offer some benefit simply by not running (and getting thrown out). Speed isn’t everything, though. Eric Hosmer is a pretty fast player, but he’s gone first to third on a single only once out of 17 chances this year, and he’s been thrown out on the bases another three times, leading to a -5.7 UBR.

So what happens when we combine runs on the basepaths with runs from stolen bases and throw in runs on double plays?

We see a decent relationship here. A player’s speed definitely has a bearing on his baserunning. Perhaps this seems simplistic, because of course it should. It’s not a tremendously strong relationship and is an indicator that non-speed skills like getting a good jump and making a good decision are likely very important as well. Speed matters, and perhaps one could rank fastest players by sight and come up with a similar conclusion, but it’s nice to know when relying on Sprint Speed that it does seem to show up in stats that matter. Going further, we might want to look at full-season stats. We might want to look at just range for outfielders. As a start, though, the data seems promising.