Stealing First: Speed and Batting Average

I recently was rereading through George Will's Men At Work: The Craft of Baseball for the umpteenth time. If you haven't read it, I recommend it. Even though it's 20 years old, it's still a fascinating look at the game. As I was saying, I was skimming through the book, and found this interesting quote...

Reggie Jackson, who can not be accused of under-emphasizing the importance of power, is mightily impressed by the sort of speed possessed by Vince Coleman. Jackson notes that in 1988, 34 of Coleman's 160 hits never left the infield — they were infield hits and hits on bunts. Take those away and Coleman would have hit .205 instead of .260. "So he ran .055 points. He's got to hit .200 and run .100 to bat .300."

Now, Coleman's speed is something that this game has rarely seen. That sort of speed can entirely alter a game. It allowed him to essentially "steal first" on infield hits that should have been outs several times over the course of a season. So, despite what Tim Raines says, maybe it is possible to steal first base, at least occasionally.

It got me wondering, what players in recent history have benefited the most from their speed in the realm of batting average. In addition, what would their numbers have been like had they been an average runner? And just for the fun of it, who would've benefited the most from being a bit quicker?

So to start looking at this, I gathered all the data on infield hits, bunt hits, etc. for qualified players since 2002 (The first year FanGraphs had all the data needed). From there, I calculated the year-by-year league averages (For the qualified players) of infield hit percentage and bunt hit percentage. Possibly the biggest surprise to me of this whole exercise was that the league average bunt hit percentage was consistently hovering around 30%.

At this point, all that was needed was to get the expected number of bunt and infield hits for each player assuming he was an average runner. This is just the rounded off infield hit/bunt league average times their number of opportunities for each scenario. Below are the top 25 individual player-seasons who benefited from their speed.

Many expected names on that list. The only potential surprise could be Jason Bay's 2008 season. His 10 SB that year were the lowest of the players on the list. For other surprises from the top 50, try 2007 Jose Guillen and 2006 Jose Lopez (Both with 0.022 Diff and 5 SB). For reference sake, the league average Diff was -0.0002, explained by the rounding error involved in calculating Exp Avg.

Does this mean that Ichiro was really a .280 hitter all those great years? Of course not. It's just interesting to see how speed can really benefit a batting average. And now, just for fun, let's see the batters whose running hurt them the most. Or, "The guys who run like an ice truck."

Lots of catchers and DH-types. No real surprises there either, though Placido Polanco did go 14 for 16 in SB attempts in his year. Other surprises in the bottom-50 include 2002 Orlando Cabrera (25 SB, -0.016 Diff) and 2003 Brian Roberts (23 SB, -0.015 Diff).

Now, some of these comparisons aren't entirely fair. I mean, David Ortiz wouldn't be expected to beat out many infield grounders because he's not built to do so. He's built for power. But this still can bring up interesting questions for players. What's going to happen to Jose Reyes (0.0048 Diff in 2012) as he ages? He probably won't beat out as many infield hits, and his average could potentially suffer.

Ichiro could be a potential example of this. You'll notice 7 of his 11 seasons in the dataset land in the top 25 of Diff. What were his two worst seasons for Diff? 2011 and 2012, when he posted Diffs of 0.0177 and 0.0127 with batting averages of .271 and .283. His Diffs still highly respectable compared to the rest of the league, but low by his lofty standards. And his batting averages were by far the two worst of his career. It's entirely possible that his Diff number will continue to creep toward 0 or worse as he ages, and think where the average could go if that happens.

This Diff measure could be used to look at a few different questions: Whether lefties really have an advantage coming out of the box on infield hits. How much a player's speed diminishes with age. It would probably be possible to regress age vs. Diff to try to predict a player's batting average as he ages. But for the moment, it's just meant to be an interesting look at who benefits the most from their speed.