On BABIP, Small Samples, and Overreactions

Success in sports wagering requires you to use knowledge that the average bettor does not have. Unfortunately, most of us do not have access to advance injury reports or friends who are trainers for professional teams. Statistics are the only tool we have to gain an inside edge; or, more accurately, knowing which statistics are useful and which are misleading.

The early weeks of the season present a great opportunity to take advantage of Average Joe Bettor's overreaction to early season trends. Often times a pitcher will get "hit hard" a few times early or throw a couple of gems, and the public will decide that that pitcher is off to a rough start, or a hot start, and the lines may adjust accordingly. But more often than not, the pitchers' results will regress to their usual numbers by season's end.

How can you tell if a pitcher has really improved or regressed, or whether they've simply been lucky or unlucky in their first few starts? A useful tool is BABIP, or Batting Average on Balls in Play. As you may know, a few years ago, a guy by the name of Voros McCracken did a study of pitching performance and found, for the most part, that pitchers have very little control over whether balls put in play result in outs or hits. Pretty much the only significant things a pitcher can control, McCracken found, are walks, strikeouts, and homers. The article can be seen here: http://www.baseballprospectus.com/article.php?articleid=878

The article has been validated by numerous studies since then, although there are a number of "caveats" for knuckleballers, extreme fly ball and groundball pitchers, and few other things. These exceptions are not important here. The larger point is that by looking at a pitcher's BABIP over his first few starts, you can figure out if he's truly improved or regressed, or has simply been lucky or unlucky. BABIP usually averages out to something in the .290s. If a pitcher has an early season BABIP well above that, his early-season troubles may well prove irrelevant. If he has a very low BABIP, his hot start may turn out to be nothing more than a couple of fortunate outings.

Let's use one example from last week. Prior to his most recent start on April 16, Roy Oswalt was generally perceived to be struggling mightily this year. He was 0-3 with a preposterously high ERA. Due in part to these preceived struggles (as well as past issues with the Phillies), Oswalt was actually an underdog going against the Phillies, who were sending Kyle Kendrick to the mound. This only makes sense if you assume, as many betters did, that 2008 Oswalt was not the Oswalt we all know and love. But a quick look at his numbers showed us that Oswalt's horrid start was in part due to bad luck. He had a ridiculous BABIP somewhere in the neighborhood of .450 prior to that game. He had given up 5 HRs, but 4 came in one game- a sample size too small to be relevant. He had struck out 12 and only walked 2 in those first three games. Was he an ace through his first three games? Not really. But he was also very, very unlucky, and that played a part in making him an underdog vs. Kyle Kendrick. Read the last part of that sentence again. Last week you could get positive money on Roy Oswalt facing Kyle Kendrick. I'll take that any day of the week. And I did. Oswalt went seven strong innings, allowed just one run, and the Astros won despite a surprisingly strong start for Kendrick.

It appears that we only had a one-game grace period before the public caught up, as Oswalt is starting today against the Padres and is -140 at last check. But it pays to keep your eyes peeled for "hot" or "cold" starts, to use BABIP to figure out whether they're really hot or cold, or just a matter of pitching luck.