Back on August 12 I wrote a piece claiming that Alex Avila‘s struggles at the plate could be blamed, at least in part, by an “unlucky” BABIP. My thesis was that Alex’s batted ball types (line drive rate, ground ball rate, etc.) predict a much higher BABIP than he was seeing (0.331 instead of 0.261). I calculated that if his batted balls fell in at the expected rate his slash line of .212/.300/.310 would have instead been approximately .269/.348/.375. The conclusion was that we should expect Alex’s seasonal rate stats to improve over the final month and a half of the year. I remembered on Friday that I never checked to see if that prediction came true or not.
My first check was to determine his xBABIP (expected BABIP) in post August 12 games. I wanted to make sure that he was still hitting the ball more or less the same as he always did so that our comparison was apples to apples. Using the same xBABIP calculator, I came up with 0.334, which compares very well to the 0.331 number from earlier in the year. Let’s call that a wash.
Now I was ready to look at what type of numbers Alex put up in the 37 games he played after my original post. His actual BABIP was 0.305. That’s still a bit lower (quite a bit actually) than the calculator predicts, but it’s miles better than the 0.261 he posted prior to August 12. The result of this was a much improved slash line of .255/.344/.391. That’s pretty close to, and actually a bit better than, my prediction.
So what are the takeaways? First, this is how we should be using BABIP, especially for hitters. We should always be comparing it to an expected value. It’s not a stat that should be discussed in a vacuum; there’s always context involved. For young hitters, an xBABIP calculator is a handy tool, but career BABIP numbers probably work just as well (if not better) for expected numbers for veteran players (plus they’re a lot easier to come by).
Secondly, BABIP and random variation are not just examples saberist mumbo-jumbo. We just showed that it was reasonable to predict an increase of over .100 points of OPS based solely on an expected increase of BABIP. We shouldn’t get too down on guys that are going through stretches of poor luck* (and we shouldn’t get too high on guys going through stretches of fantastic luck. See: Boesch, Brennan). I wouldn’t call a few points difference (maybe even ten or so, I don’t know) between BABIP and xBABIP lucky or unlucky, but the 0.070 we saw from Avila? That sounds the alarm bells.
Finally, BABIP doesn’t explain everything, and BABIP is not totally explained by randomness. There are other non-BABIP factors that aided Alex’s improvement. An easy one to see is that he hit three of his seven home runs during the final stretch. Home runs are not included in BABIP, but it surely helped out his slugging percentage (which ended up well above our prediction). Is it reasonable to expect Alex’s BABIP to climb all the way to 0.331? I’m not sure yet. That’s what the calculator says, but it takes in general data and spits out a number that probably works for the average player; that’s why I like to use career numbers for veteran players.
Alex finished the year with an OPS of 0.656, but I don’t think it’s unreasonable to expect that number to be in mid .700’s next year.
*Any time I use the word “luck” I really mean random variation. That’s really how I’m using the term, not in the sense of the literal definition of the word. I don’t believe that type of luck exists.