Baserunning is not a very big part of player value. In 2018, only one player gained more than a win thanks to his baserunning, and no player lost more than a win. A quick glance indicates that only around 20 percent of full time players even have their WAR impacted by more than half a win as a result of baserunning. So, it’s something that mostly helps (or hurts) at the margin. But that doesn’t mean we should ignore it.

So, what all goes into baserunning? Well, it’s three pretty simple things. The first should be obvious: stolen bases. Remember how we talked about linear weights for hitting? Well, you can do the same thing with stolen bases (or caught stealings) as for other offensive outcomes. What’s the run value of moving up a base? What’s the run value of getting thrown out? Calculate the average such value across the landscape of all possible situations (which, recall, includes both any runs that score as a result of this, and any runs that score/don’t score after it happens), and there are your run values for stolen bases and caught stealings. Take those values, apply them to the actual number of steals and caught stealings by a player, done. Blam.

The second is taking extra bases during balls in play. In practice, this is very annoying to calculate. But in theory, it’s not too different from how arm metrics are calculated for outfielders, combined with the linear weights methodology of valuing extra bases. Using the same linear weights method, we can calculate the value of advancing an extra base in an average situation (or getting thrown out trying to take one). We also know how often runners take extra bases based on where the ball is hit. So, put those together, and you get a system where:

- First, a ball is put into play. Based on where the ball is put into play and the current base-out state, you know the
*average*result, in terms of runs (both as a result of the ball in play, and what will happen later in the inning) as a result of that. - Now, compare that
*average*result with what actually happened as a result of the runner(s) on base. Did they move more bases than average, or fewer? If fewer, that’s below average, i.e., some combination of fewer runs scored as a result of the ball in play and/or fewer runs will score later in the inning. If more, that’s above average. Naturally, outs on the bases reduce runs, while not making outs and taking extra bases tends to increase runs.

Again, thinking about doing this in practice can be kind of head-spinning. But it’s already done for you as “UBR” for each player, you don’t need to do it yourself. All you need to do is just acknowledge that this is the component that captures the value of taking extra bases and/or getting thrown out in trying to do so. And remember that if on a given ball in play, an average runner advances two bases, but a runner only advanced one, he’s losing value (but not as much value as if he had gotten thrown out).

The third component probably *could* be considered part of batting, but for the sake of theme, is included in baserunning: avoiding double plays. This measure just considers how many double play opportunities a player had, how many he hit into, and what the average rate of hitting into a double play during a valid situation is, and then just credits or debits the player accordingly based on whether he avoided hitting into double plays or did so more often than league average (again, adjusted for opportunity). The debit is essentially one extra out of run value charged to each player for each double play hit into above average, or one extra out of run value credited to each player. Of these three components, the double play one tends to be the most marginal, but there’s always a few players who either claw back half a win or so by refusing to hit the ball on the ground *and* beating out double plays when they can, and a few who toss half a win away by hitting tons of grounders despite being quite slow.

**Sprint Speed and Maybe Burst One Day**

Statcast doesn’t have its own baserunning metric. But, it does publish sprint speed for players. While baserunning metrics are specifically about value, sprint speed is just that, a measure of speed. And, while it might seem intuitive that sprint speed would correlate with the various baserunning value components, that relationship is not exactly clear-cut (see here: https://www.talkingchop.com/2018/9/18/17866544/atlanta-braves-sprint-speed-and-baserunning-value). It’s not that it’s not related, it’s just that it’s not a slam dunk:

You see the best effect in large samples, and mostly for the UBR-based “extra bases on balls in play” component; perhaps surprisingly, stolen base value isn’t that correlated with footspeed. Why is this the case? Well, one reason could be that sprint speed is measured as a top-line “max speed” figure on lengthy runs, i.e., those covering two or more bases, or on home-to-first sprints. Statcast and Baseball Savant also mention a burst metric, which is sort of like a measure of acceleration after the first step of a run is taken. Given that stolen bases usually involve a lead, and also feature a slide at the end, the highest velocity of a player (as captured by sprint speed) is probably less relevant than acceleration (burst) along a shorter path. You can check out Baseball Savant’s Sprint Speed Scroller (https://baseballsavant.mlb.com/sprint-speed-scroller) for a greater explanation; I bring this up only to say that while the Statcast data we *currently* have access to aren’t really an alternative measure of baserunning value or a great way to predict future baserunning value, they might be one day if burst or similar metrics become available.

**tl;dr takeaway for baserunning- **baserunning is a small component of value. It’s more than just stolen bases — it also involves taking extra bases on balls in play and avoiding double plays.