I started collecting play-by-play data around three years ago in order to measure a very specific phenomenon, one so important to me I can't remember what it was. I do it manually, and the data itself isn't overwhelming since I can do a full slate of 15 games in about an hour. No, it's the 80+ columns of calculations and what-not I add that increases the complexity, but out of that chaos can come very interesting data.
Very quickly I realized some near-and-dear notions to baseball purists didn't hold up to scrutiny, and since it's -15 as I write this, I'll do anything to keep my mind off the cold. I became intrigued in seeing how often a batting order slot bats in that given slot. In other words, how often does a leadoff hitter actually lead off an inning? How often does the cleanup hitter bat fourth and so on. As we get nearer to spring training we'll begin hearing things like "Team X needs a leadoff hitter" or "Team Y would be better batting so-and-so 6th," or my personal favorite "This hitter adds punch to the bottom of the lineup."
I have complete play-by-play data going back to 2009 that's easily accessible and data going back to somewhere in the mid-1980s that isn't as complete, all of which was accessed from Baseball-Reference.com and beaten to within an inch of its life by all sorts of Excel text and data extraction equations that I almost understand.
This chart is the crux of this post and shows how often a batter actually bats his batting order slot:
Click on image to enlarge
The vertical axis is a player's batting slot, 1 through 9. The horizontal axis is what spot in an inning a player actually batted. For example, the leadoff hitter actually led off an inning 40.7% of his plate appearances, the #2 hitter 41.6% of the time and so on.
Obviously, we can expect batting order positions 1 through 3 to bat their slot more often because they're guaranteed to do so in the first inning, but after that, it becomes a matter of luck. Cleanup and #5 might be worthy of thought, but after that, it's random chance at work. Somewhere in this data is probably support for batting the pitcher 8th instead of 9th, but it's beyond me as to where it is.
This data poses a question--should a hitter's approach change depending on where he's batting in an inning? When Joey Votto leads off an inning, should he approach his at-bat as a leadoff hitter, working the count and feeling the pitcher out or as the #3 hitter he typically is, looking for a pitch he can drive? Should Mike Trout approach every at-bat with a leadoff hitter mentality? Obviously, runners on base, game situation and myriad other factors will come into play, but it does make for an interesting dichotomy--is a hitter really a #3 hitter or is he what he needs to be in each plate appearance. The answer, of course, is yes.
This chart shows the number of batters per inning between 2009-2013:
37.8% of innings are three up-three down innings, 28.2% have four batters and so on. Consider this fact alone--around two-thirds of innings feature only three or four batters, which makes every base, be it gained by hit, walk, error or other manner, valuable. It would be interesting to see if these numbers vary much over time, but I suspect the change is fairly small.
The numbers appear to speak for themselves--the traditional expectations of the top three positions are well-founded since it's guaranteed to occur in the 1st inning, but after that, managers are probably better served mixing up left-handed hitters and right-handed and keeping slow runners separated than trying to maximize the value of a sixth hitter. It's already happening, since when was the last time you heard a manager say "We need to find a #6 hitter" as opposed to some blogger like me who doesn't know any better? Not often, if ever. This is the type of research I do--I don't claim to be an advanced metric guy using sophisticated statistical analysis as much as a person who wades through mounds of data looking for trends and then hopefully explaining them in a manner both informative and enjoyable. Of course, that's not for me to judge.
All data adapted from Baseball-Reference.com
Scott Lindholm is a web columnist for 670 The Score in Chicago. Follow him on Twitter at @ScottLindholm