Examining the Wizard: Roy Oswalt, a Case Study in Luck

Today, I was excited to see this article from The Hardball Times. For those of you who are too lazy to click the link and read the article, I’ll quickly summarize. The article looked at pitchers who were the unluckiest so far this season; measuring their rate stats against league averages to define unlucky. Astros fans should not be surprised to find out that Roy Oswalt was the most unlucky pitcher in terms of HR/FB (Home Runs per Fly Ball), posting an absolutely astounding 21.05% HR/FB so far this season. The article by, Derek Carty, goes on to normalize Oswalt’s (and the other pitchers listed in the article’s) outlandish rates stats back to league average levels, plug it into a very complex formula, and determine Oswalt (and company’s) LIPS ERA (Luck Independent ERA). Carty finds that by normalizing Roy Oswalt’s HR/FB back to league-average, Roy Oswalt owns a LIPS ERA of 4.08. Not an outstanding ERA, but I don’t think any Astros fan would quibble with having Roy Oswalt’s ERA be 1.30 earned runs lower.

The LIPS ERA concept works, because by-in-large, things like BABIP, HR/FB, and LD% are largely beyond the pitcher’s control (See above linked articles and this one for supporting evidence). However, what does the knowledge that Roy Oswalt has been unlucky tell Astros fans, or fantasy owners for that matter, about what to expect? Thankfully, Derek Carty again, provides some insight into a favored term bandied on this blog “the law of averages” or “regression to the mean”. Once again, for those to lazy to click the link, essentially we can expect Roy Oswalt to surrender HR/FB at a rate consistent with his career levels from here on out, meaning somewhere around 9% HR/FB, ( he’s never exceeded 12.9%).

To dig further than just discussions of luck and regression to the mean, I was graciously given Roy Oswalt’s pitch-result data from the talented and majestic Josh Kalk. Honestly, I cannot thank him enough for helping me, who knows absolutely nothing about database coding, for providing this data. While I had just about anything you’d ever want to know (and a lot you wouldn’t about) Roy Oswalt’s every pitch in 2008; save his last two starts (tonight’s start included) due to when I actually obtained the info. I’ve isolated only the pitches that resulted in HR, since that’s what we’re looking at (click here for an explanation of the data):



Quickly, here are the pitch averages for all of the available data on Roy Oswalt for 2007:



And then his 2008 pitch averages:



Clearly, movement is down on his Fastball (FB) and slider this year. Velocity does not seem to be the issue for Roy on his fastball in terms of HR allowed either. Of the 11 FB’s that turned into HR in 2008 for Roy, the average speed was 93.23 MPH (min: 91.62 max: 95.61), so that’s clearly not the issues as evident by just looking at his 2007 average. His movement on the FB’s that turned into HR: -4.09 in. X (min: -1.44 in. max: -8.16); 6.00 in. Y (min: 1.99 in. max: 10.85). There’s not as much life on the FB’s that became HR this year by comparison to Roy Oswalt’s 2007 numbers and on the whole his movement is just down. This probably explains his overall decline in strikeout related rate stats, but I don’t know that it really explains why he’s just getting crushed so much. However, I get the feeling that 11 FB's is just far to small of a sample size to say anything meaningful about. In fact if anyone can see any clear trait from the FB velocity or movement from the chart that indicates he's doing something that's allowing his HR spike, please speak up.

Note: I’m ignoring the three sliders, because his slider is just God awful this year, but not the issue apparently.

In my last attempt to decipher the enigma that is Roy Oswalt in 2008, I cited the fact that his release point seems excessively bunched, most likely in an effort not to tip his pitches, thus causing him to lose life on his pitches. While my initial conclusion from that observation was that it must be his slider that is getting pounded was wrong, I none the less stand by the observation as providing insight into his struggles to date.

Tonight, however, I am offering another observation. Looking back at Roy Oswalt’s HR pitches, all of but three of the fourteen pitches have one thing in common: men on base. While it is clearly a limited sample size, perhaps either pitching from the stretch, or the distraction of holding runners is causing Roy to throw pitches that are a bit more crushable. Food for thought at least. While it’s not an earth shattering observation, the pitch/fx data of his HR pitches revealed no other apparent trend. Oddly though, his 50/50 split of RHB/LHB going yard is elevated for his career 75/53 RHB/LHB HR split. There wasn’t a meaningful home/road split for the HR’s either.

Having spent the entirety of the game researching and writing this article, we saw some dominant pitching from Roy Oswalt tonight, but the Ryan Braun long ball makes it 17 HR the season. During the post game so far I’ve heard a lot of discussion centering around Roy Oswalt trying to pitch to contact more vs. trying to strike guys out. Our first attempt to discern what was ailing Roy Oswalt revealed little in the way of evidence that Roy was attempting to pitch to contact from his numbers, but there are probably short comings in trying to prove or disprove pitching philosophy from limited statistics. After an exhaustive combing of pitch/fx data, I’m left scratching my head. It seems that although pitch/fx is an incredible tool that is sure to further revolutionize baseball analysis, it also seems to support the assertion that pitchers have little control over their BABIP, DIPS, HR/FB, LD%, etc. If this is the case, we can only have one logical expectation of Roy Oswalt from here on out: that he should pitch in-line with his career numbers. Astros fans should rejoice and fantasy owners should do what they can to acquire him.

Sphere: Related Content