clock menu more-arrow no yes

Filed under:

The Braves, sprint speed, and baserunning value

New, 51 comments

The Braves are fast. But how much does that really help them?

Tampa Bay Rays v Atlanta Braves Photo by Kevin C. Cox/Getty Images

The 2018 Braves are fast. Not all of them, certainly (hi, Tyler Flowers and Kurt Suzuki!), but as a team, they’ve got a lot of speed.

Source: Baseball Savant

Those colored dots are the Braves, and the gray dots are everyone else. Sure, the Braves aren’t above-average everywhere, but they’re above-average enough. (Nick Markakis and Johan Camargo aren’t much help, though.)

Looking over these numbers, though, might lead you to wonder: “Okay, the Braves are pretty fast. But how valuable is that?” Luckily, baserunning value is tracked and accounted for. A quick glance at the current Fangraphs leaderboards will tell you that the Braves are seventh in MLB in baserunning value, 8.1 runs above average. Now, for some of you, that might be satisfying. The chart above shows that the Braves are pretty fast. The leaderboard shows you that they’re pretty good at accumulating baserunning value. A decent stopping point.

I had a different question, though. Actually, two questions. First, “How fast is pretty fast?” And second, “For how fast the Braves are, what should their baserunning value be?” The rest of this post is an attempt to answer both of those questions.

How fast are the 2018 Braves?

When we talk about how well a team hits, what we’re really just talking about is a weighted average of the hitting abilities of the different players on that team. To the extent a player hits more often, his abilities are weighted more heavily in the average. To that end, we can do exactly the same thing with sprint speed. But, there’s one wrinkle: a player’s hitting value is tied to how many PAs he gets. But, for baserunning, that’s not quite the case — a player can’t accumulate baserunning value if he’s not actually on base. So, there are two ways to think about a team-level, weighted average sprint speed: (1) just a straight weighted average by PAs; and (2) a weighted average of PAs multiplied by OBP, to give those players on base more often a greater weight to reflect their greater share of baserunning opportunities. In the end, these two ways don’t matter very much, because OBPs don’t vary that much. The table below shows where the Braves rank, along with every other team, by weighted average sprint speed.

So, if you thought the Braves were pretty fast, your instincts were correct. They’re the fifth-fastest when doing a weighted average with only PAs, and eighth-fastest when weighting by the product of PAs and OBP. Done with the first question.

What should the Braves’ baserunning value be?

This question, well, it’s tricky. First, we need to talk about the components of baserunning. Specifically, there are three of them, as captured by Fangraphs and recorded for each batter.

  1. UBR. UBR, which stands for “Ultimate Base Running” (yes, really) (and is not an extreme sports competition featuring Billy Hamilton and Byron Buxton), measures the value provided by baserunners taking extra bases. This is actually relatively easy to measure and define. If you’re interested, you can read more about it here and here. Speed presumably plays a large in UBR: it’s easier to advance safely if you’re faster. However, UBR can be a bit tricky, because the opportunity to advance for Player X is generally predicated on Player Y putting a specific type of ball in play. To that end, players who hit/reach base in front of players with non-average batted ball profiles may have their UBR look weirder, for a given sprint speed, than other players. In any case, we would expect faster teams/players to have higher UBR scores.
  2. wGDP. This is basically a measure of how well a player does/does not avoid double plays. In some ways, this isn’t speed-related at all: if you hit fewer grounders, you will hit into fewer double plays. But, faster hitters do tend to avoid double plays more. wGDP doesn’t generally matter much: the spread between the best and worse baserunner in terms of UBR tends to be around 10 runs, or one win. It tends to be smaller for wGDP. However, Braves fans should be intimately familiar of when wGDP can be a real killer due to having to bear witness to Matt Kemp’s 2017 season, where Kemp was a bottom 50 runner and a “top” 50 groundball rate hitter, leading to one of the worst per-PA wGDPs ever. In any case, we would potentially expect faster teams/players to have higher wGDP scores, but this relationship should be expected to be fairly weak because it depends a lot on groundball tendencies in addition to footspeed.
  3. wSB. This is probably the most obvious thing that people think of when they hear “baserunning,” i.e., stolen bases, and not getting thrown out on the attempt. The spread of wSB also tends to be relatively low (similar to wGDP), though this is largely because teams don’t run much. The spread was bigger, similar to wUBR, before the juiced ball era. Of the three statistics, you might expect this one to mesh with sprint speed the most. Sure, fundamentals like a good crossover step and an understanding of when and when not to run matter, but the result here is largely within the runner’s control, as opposed to UBR (because it relies on getting a ball in play when on base that allows for running to happen) and wGDP (because it relies on not hitting grounders, which happens before the runner reaches base).

The thing is, I just wrote three paragraphs about why we think sprint speed should be related to these measures... but in reality, the link isn’t great.

Actually, “isn’t great” isn’t really a good descriptor. Nonexistent, more like. The red dot is the Braves on each chart, and you can see that the Braves are generally pretty good at UBR (and overall baserunning), but not so much wGDP or wSB. Note that doing alternative versions of these scatters using the PA x OBP weighting, or calculating the baserunning values on a per-PA basis (because team PA totals differ) does not change this at all.

Given this lack of correlation, we’re actually not too close to answering the question posed. If, on a team basis, sprint speed doesn’t track with baserunning value, then it’s hard to say it should be this or that. A simplistic analysis could say something like, “Well the Braves have the fifth-highest sprint speed and the fourth-highest UBR, so that makes sense. They have the fifth-highest sprint speed but only the 12th-highest wGDP, but even that’s not worrisome because they hit their fair share of grounders. But, what’s the deal with their fifth-highest sprint speed, but a wSB of -0.6, only 16th in MLB? That’s the part that doesn’t add up.” And yeah, that’s simplistic, because as the scatterplots showed, perhaps we shouldn’t expect these ranks to line up.

So, let’s look a little deeper.

The scatterplots above plotted the various stats on a team basis. But, what if we did the correlations on a player basis instead? I’m not going to post a bunch of scatterplots which mostly show the same stuff, but here’s the basic idea: the columns are the different baserunning components, and the rows are different “cuts” of the player universe. What you get is something like this:

To illustrate this with just one scatter, using BsR (the aggregate measure) and the per-PA basis for the sample of only players with 400 PAs or more, the relationship looks like this:

I added the equation on there only to facilitate the dubious math I’m about to do next. Based on this relationship, we expect that each foot per second of sprint speed should add about two runs per 600 PAs of BsR value. How does this actually compare to the Braves and what they’ve done?

I find this table pretty interesting. In general, the one-run fluctuations aren’t really a huge deal, they’re more of a curiosity than anything else. But, there are a few differences I find kind of notable. For one, we know that Ender Inciarte isn’t fast — see the very first image of this post, where he trails most center fielders in footspeed. Yet, despite that, he’s actually racked up a fair bit of baserunning value. Johan Camargo is the opposite — he’s slow, but has been much worse on the basis than his slowness suggests. The part that gets really striking to me, though, has to do with wSB. In short, if your wSB is negative, you’re probably better off not attempting steals. That Johan Camargo has the biggest wSB negative is not surprising in light of the other columns: he apparently makes decisions like he’s fast, but he’s not. But, what’s Ender Inciarte doing with a negative wSB?

Inciarte appears to be in kind of a strange limbo here — he runs the bases well, especially relative to his speed. But while his 25 steals lead the team, he’s been caught stealing more times than the next three Braves combined. Inciarte has stolen 20 or so bases each year of his career, but his wSB has generally hovered around zero (except for his rookie season). This is the first time in his career that it’s dipped below zero, suggesting that there’s some room for improvement in terms of decision-making.

But, overall, this probably isn’t an area the Braves need to focus on too much in the offseason. Not only is baserunning a somewhat fringe component of offensive value, but the Braves don’t seem to be doing too badly, considering their footspeed or no. Just figure out a way to relay to Johan Camargo that he’s not actually a speedy middle infielder, and that could be enough.