OK, I loved this post. Great job. It's got all the things I love (graphs!). I think your comment above to OGC is well stated >> I certainly don’t think of this as a big blow against DIPS. To me it looks like the ability to induce pop-ups is along the lines of owning a sweet pickoff move. It’s a nice extra advantage, but not an ability that I would pay money for. It seems like, to me, that pop-up ability is a little bit like baserunning (though, maybe not the same level of impact): it's a nice little bonus skill to stick in your back pocket, but by no means a 'meat-and-potato' skill. I had a side question about Zito's pop-up skills...how much did Zito's park in Oakland influence his abilities to make hitters pop-up? He posted some truly huge infield flyball rates while in Oakland and from my recollection of park factors, the A's stadium has always been above average for infield fly balls. I'm not asking you to park adjust anything, but it's an interesting side thought.

Nice work, Otis!

Okay, I'll cram all the technical stuff in a nice skipable comment. The graph is a graph of density curves. You may be familiar with the bell curve (or normal distribution). That's also a density curve. The y-axis basically tells you the fraction of the results that are to the left or to the right of a given point on the curve. To left of the mean, it tells you the percentage of the group with that score or lower, and to the right of the mean it's percentage with that score or higher. I normalized all of the ball in play totals so that they could be compared to each other. I measured the mean of each ability, and calculated a z score for each pitcher using the results that each pitcher achieved, the results he would have achieved if he were just average, and the standard deviation appropriate to the number of opportunities he had to achieve each result. The density curves are all plotted from the z scores. I used R-studio to do most of the work, which I can't recommend enough.

Very nice, I've been wondering about this myself but didn't have access to the data to explore this. And thanks for the hat tip to R-studio, I will have to check out one day. First, it would have been nice to include the number above average for each pitcher in your table. I was unsure what the ranking function was until your text later. Second, I think providing another table showing the ranking based on number above average divided by total flyballs would have been nice to see as well. As your current table makes clear, there are those much more skilled on a per flyball basis, and I wonder where Zito (and Cain) would rank when based on that. Of course, small sample sizes affect that, but still would be interesting to see. Now there are hypothesis testing one could do, right?, to check to see whether the null hypothesis holds or not, if I remember my stats right. Couldn't that be used to test to see if any of these guys are significantly above zero (i.e., the mean). Just a thought towards future analysis you could do. What I'm leading to with the above, is that I think Matt Cain is also a popup artist like Zito. But the way you are presenting the data, because he hasn't pitched for a good portion of the 2001-2010 study period, nearly a half, he is handicapped by that, whereas Zito was a starter the entire period. Just doubling his results, to put his career IP closer to Zito's, would put him at 60 popups for the period, which would be pretty close to Zito's 79. Great stuff, I look forward to your future analysis in this area. I think that this is a huge piece of the puzzle for explaining why Barry Zito and Matt Cain have more value than current saber-zealots who have DIPS-blinders would acknowledge. I think that the current state of art places pitchers like them in the same regard as hitters with low batting averages but good walk rates prior to the 1990's or even pre-Moneyball: unappreciated by the masses. You would think that people would remember the hubris from thinking BA was it, then the fall realizing that OBP was more valuable. OBP was not really appreciated until then, and I think pitchers who are capable of inducing popups at a higher rate than others fit into that unappreciated category that further data and analysis will uncover. Nice start to uncovering this area.

This post is filed under: stuff I cannot do... or at least don't know how to.