Infielders vs. outfielders: Taking the extra base

Few things define grit and hustle like taking the extra base. Going from first to third or second to home on a single and going from first to home on a double are things that get a lot of praise from managers, teammates, and broadcasters. Really, fans should be impressed too. Running's hard, y'all, but outrunning the expected base state is like outrunning the future. The future stands at third base on a double; savvy and/or speedy baserunners are heading to the post-game spread after scoring the game-winning run on a double.

Ian Kinsler, an infielder, is often mentioned as one of those savvy baserunners. Indeed, he is pretty good at taking the extra base; his extra-base rate is right next to noted speedster and CF-defense savant Kevin Kiermaier. Though their overall extra-base rates are about the same, Kinsler is much better at taking the extra base on a double or when he stands on second when a single is hit. Kiermaier is much better at going first to third on a single. Some veteran savvy at work.

For reference, the following data come from multiple sources and are 2015 only. Baseball-reference has the extra-base data on a player level, but keep in mind that the "location or type of the ball in play" is not taken into account. I then grabbed each player's Speed Score (which I am using as a proxy for actual speed) and position from FanGraphs. To get a player's position, I grabbed each player's innings defended by position and assigned the "primary" position as the one with the most innings. Most players don't split time between the infield and outfield enough to make that rough method inappropriate, so it should be OK (unless you're Ben Zobrist, though he spend more time in the infield this year). I also grouped 1B, 2B, SS, and 3B as 'IF' and LF, CF, and RF as 'OF'.

I came into this with a bit of a hypothesis, which may or may not seem idiotic to you. I thought that outfielders, given their outfield experience, would have a better idea than infielders of when to take the extra base. Of course, there is a wide range of body types and speeds across all the positions, so I attempt to use the Speed Score to control for that. I also grabbed the age of each player in case that was a confounder.

There are four metrics I'll look at: overall extra-base rate (which is a composite of the other three), first-to-third rate on a single, second-to-home rate on a single, and first-to-home rate on a double.

Comparing the groups, outfielders come out on top in every metric. The age difference is minimal, but outfielders are indeed a little faster. Since they have a bit more speed, I would expect them to have better rates. Makes sense. Infielders include lumbering first basemen, who are lumbering and not fast.

There is indeed a way to make the groups a little more comparable in terms of speed. Excluding the corner infielders yields equal speed scores, but the age difference is a little larger. When comparing 2B/SS vs. LF/CF/RF, the infielders come out on top.

Does that seem fair to you? Though the average speed scores are equal, there are still some lumbering guys out there in right field. However, excluding right fielders gives the outfielders the speed advantage once again. Despite that speed advantage, the outfielders don't take advantage of it.

What about trying to compare more "like for like"? How about the speedsters of the outfield, the center fielders, vs. the speedsters of the infield, the shortstops? It's an interesting result.

Despite a nearly 31 percent higher average speed score, the center fielders don't dominate the shortstops in the metrics. Hmm. To me, that suggests that speed has diminishing returns. After all, there is only so much speed you can utilize in moving in a defined distance, and humans can move only so fast.

As a side note I would like to see Terrance Gore vs. Yoenis Cespedes in this vein. It could just be a weird thing in the data, but I wanted to investigate anyway. Thus, I fit two trend lines to a scatterplot of speed score vs. extra-base rate. One was a linear fit, and the other was a logarithmic fit (restricted the scatterplots here to only those players with at least 100 PA).

The linear trend model describes the data just a little bit better, but I kind of feel like the logarithmic equation does a better job at the high end of speed scores, while the linear model does a better job at the low end. I might just be seeing what I want to see.

Aside from the trend models, Holy Trayce Thompson, Batman! Don't read too much into it; he's had all of eight total opportunities to take the extra base on a hit. That's a pretty small sample size.

At the end here, there might a few takeaways. First, infielders as a whole group are worse at taking the extra base than outfielders due to lumbering first basemen. Second, the faster infielders might be better than outfielders at taking the extra base in that they seem to be able to do more with a lesser speed score. Third, there is a slight hint of diminishing returns to speed.

. . .

Kevin Ruprecht is the Managing Editor of Beyond the Box Score. He also writes at Royals Review. You can follow him on Twitter at @KevinRuprecht.