Incredible Oscillating BABIPs

There’s been a lot of work done on BABIP by the esteemed writers on this site as well as The Hardball Times, so it’s not suprising that Derek Carty’s July 10th release of a simple expected BABIP calculator flew under the radar a bit. It shouldn’t have. The calculator allows us fantasy types to say with a little more certainty what sort of BABIP a particular player should expect.

The ‘simple’ in the calculator meant that instead of using the more complicated and nuanced four-component speed score, users have to use the more brute method of inputting stolen bases. This may be a considerable flaw, allbeit an easily corrected one. Taking a look at this threesome will provide us some questions about BABIP in general. It certainly looks like sometimes rotten luck lasts a whole year.

Jose Lopez – Midway through June, Lopez was languishing was a sub-.250 batting average and little power. Most were questioning if his breakout 2008 was just a fluke. His BABIP, though, was unnaturally low, and the balls began to bounce his way. He ended June with a .329/.325/.592 slash line that seemed to be a harbinger of further success. Unfortunately, his July slash line (.269/.299/.441) has been a lot less inspiring. The simple xBABIP calculator pegs his expected BABIP at .301, which seems fine at first glance. He has an average line drive percentage (18.1%) and profiles mostly as he did last year during his great year. But then look at his yearly BABIP totals, and something looks fishy. Since 2005, those BABIPs have read: .276, .312, .269, .311, .266. It seems that he has good and bad years as much as he has good and bad months. Perhaps a good year is just one where the number of good months outweigh the bad months. And maybe 2009 is ‘just a bad year.’

Vernon Wells – Wells rode a .311/.329/.514 July back into respectability and is now on pace for a career high in stolen bases. Despite his low 14.9% line drive percentage, the xBABIP calculator likes him for a .302 BABIP. In the face of his .294 career BABIP and his career-best four-component speed score, this xBABIP seems more correct than his current .278 number. Maybe the title of the article gave you a heads up, but Wells is the proud poppa of his own pendulum-like BABIP. Here are his yearly totals, since his first full year in 2002: .288, .322, .286, .275, .313, .265, .299, .278. I’m not really suggesting that that Lopez and Wells are ‘due’ for bad years, but it does seem to follow that these are two players whose value is often tied up into BABIPs, and who often put together full years with sub-par BABIPs. Caution, at the very least, is in order when thinking of buying low in these cases.

Ichiro Suzuki – Suzuki is merely here because of tendency to ‘break’ BABIP predictors. His yearly BABIP totals oscillate a little, too, but there’s hardly a year that can be characterized as ‘poor:’ .371, .347, .333, .401, .319, .350, .390, .337, .387. When .319 is a poor BABIP for you, you don’t belong in this trio. The simple calculator can’t quite handle Ichiro’s legendary BABIP control (if such a thing exists), predicting a .328 number for Ichiro which would be the second-worst of his career. With his history, it doesn’t seem smart to predict a dive in his BABIP to get to that .328 number.

The lesson seems to be that there is still work to be done on xBABIP, and I doubt anyone would argue that they’ve figured it out completely. The work to date has been impressive, but there certainly seems to be more we can understand. Why is it that some players have BABIPs that normalize within a year, and other players seem to have BABIPs that only normalize over an entire career? It may only be a question of sample size, but there seems to be something more here. At first glance, I’d have to think that contact rate could be included in any expected BABIP model, but I’ve only just begun my journey into the numbers.

Print This Post

Graphs: Baseball, Roto, Beer, brats (OK, no graphs for that...yet), repeat. Follow him on Twitter @enosarris.

7 Responses to “Incredible Oscillating BABIPs”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. lexomatic says:

    what about home/road splits for wells this year? to small a sample or craziness.

    Vote -1 Vote +1

  2. Mario says:

    Gives me a better idea of what to expect w/ Jose Lopez. Good read

    Vote -1 Vote +1

  3. Eno Sarris says:

    I think one of the things I’m saying about BABIP here is that the relevant sample size may be huge. It certainly doesn’t seem to stay very constant with these three players. Cutting it down to 2/3 of a season’s worth of splits (or 1/3 of a season’s worth of plate appearances) would be going the wrong way on BABIP I think.

    Vote -1 Vote +1

    • acerimusdux says:


      This has always been the difficulty with BABIP. For pitchers, this is the reason some early studies found no control over BABIP. Looking at year to year data, it is almost all noise. Look at careers, and you do find some significant control.

      For pitchers, it has since been found that the standard deviation in true talent ability to control BABIP is around 10 points. So if the average is around .298, 2/3 of pitchers will be from .288-.308 in true ability. But the variance in year to year samples will be much greater than that.

      Hitters have a bit more control, with a standard deviation in true talent likely almost 20 points. The probability of being 3 SD from the mean is 1 in 1000. A guy like Ichiro may be one of those one in one thousand players, with a true talent BABIP ability of near .360.

      But it’s tough to identify a guy like that for certain, especially using small samples. So any estimator will likely err on the side of assuming he’s not the one in one thousand, until there is really sufficient data to think otherwise.

      This one doesn’t seem to work for a large sample, The SB formula seems designed for single season numbers, and probably needs to be adjusted to work per AB or PA.

      Vote -1 Vote +1

  4. Chris says:

    Nice notification on the calculator…but I wish to use this time to bring up an argument completely unrelated to baseball. From an anagram perspective, BABIP is just too damn long. I like my baseball anagrams to sound like a ruthless government agency, not an excessively wordy international relief fund. Can’t we just shorten it to AIP…Average In Play? We can at least get rid of that second B can’t we? I mean, it’s not as if players are hitting chickens or orphans into play…of course they are hitting balls, that’s what they do. I don’t know…I’m just saying, that stat itself is never going to be more than four digits…the name should be four digits or fewer too.

    Vote -1 Vote +1