Interview with Derek Carty from The Hardball Times-Part 2

Here is part 2 of the 4-part interview with Derek Carty from The Hardball Times:

Fake Teams: I see you are not a fan of FIP for evaluating starting pitchers and like LIPS instead. Where does one find the LIPS calculated? I haven't seen it published as widely as FIPS.

Derek Carty: Unfortunately, LIPS isn't available anywhere publically - not right now, anyway. It was published in HEATER Magazine this past season for their subscribers, and occasionally I'll post a leaderboard at THTF, but we don't currently have it on our stat pages. There is some good news on this front, though, for those looking for LIPS. There are preliminary plans in the works to sell a StatPack, of sorts, at THT Fantasy later this off-season. This StatPack will include scores of interesting data for fantasy owners to peruse and use for their draft preparations. I'm planning to include a lot of my homemade stats (some of which I'll discuss a bit later in this interview), LIPS, and a bunch of other goodies. I'm hoping this will be an invaluable tool for fantasy owners.

Back on topic, you're absolutely right about me: I am 100% opposed to FIP for fantasy analysis. I believe you're referring to this article that I wrote, but to summarize my stance, FIP uses a pitcher's actual HRs- a huge ‘no-no'. Analysis has found (not even all that recently) that a pitcher has very little control over his HR/FB rate. He does have a lot of control over how many balls get hit on the ground (which obviously can't become HRs), but once a ball is put in the air, the pitcher has relatively little control over whether it clears the fences for a HR. FIP doesn't account for this, making it extremely inaccurate for pitchers with abnormal HR/FB rates and merely somewhat inaccurate for everyone else. While I understand LIPS isn't readily available right now, xFIP is, and this at least adjusts for HR/FB, if not for everything that LIPS does. Baseball Prospectus's QERA is also a solid choice. I can't help but cringe every time I see a statistically-inclined site mention FIP in regard to predicting future performance or in discussing luck. There's nothing unlucky about a pitcher with a 5.00 ERA and a 4.00 FIP if his HR/FB is 4%.

More after the jump: 

Fake Teams: Recently I have written a few articles on what stats help one to predict wins for a starting pitcher, and realized there really is not one single stat that does it. What stats do you look at when targeting starting pitchers?

Derek Carty: To be completely honest with you, I ignore Wins entirely. They are so riddled with noise that I don't bother with them at all. They don't even get a glance from me. If I'm looking to gauge a pitcher's win potential, I'll check out an expected Win stat I've whipped up, comprised mostly of the skills of the pitcher and the quality of his offense. As far as his skills go, strikeout rate, walk rate, and groundball rate are the three biggies. They are under the pitcher's control much more than other stats are. This probably isn't big news to your readers, however, who I imagine are pretty familiar with DIPS Theory (if not, feel free to shoot me an e-mail. I'd be happy to discuss).

As I said earlier, though, I'm not always content with sticking to what's available - sometimes I like to take things a step or two further. Last year, I did this for pitchers with the ‘creation', if you will, of CAPS (Context Adjusted Pitching Statistics). CAPS still focuses on defense-independent stats like Ks, BBs, and GBs, but it seeks to remove even more noise from the equation - from the peripheral stats themselves. For example, a pitcher facing a disproportionate amount of Adam Dunn-type hitters will strikeout and walk a disproportionate number of batters and allow too many flyballs and home runs. Because the act of facing these kinds of batters isn't a skill of the pitcher's, we should adjust for it. After all, his mix of batters going forward is likely to be randomized, not marred with 3TO hitters like Dunn. Last off-season, I rigged up CAPS to adjust for home ballpark, mix of road ballparks, quality of opposition, and league changes. CAPS totally nailed Javier Vazquez - who was critical to my LABR strategy and success - among a few others, and I believe the logic and reasoning behind it is quite sound. Without divulging too much, this off-season I plan on bringing several more context adjustments into the mix as well as creating a CAPS of sorts for hitters (CABS, I suppose it'll be called). I'm currently in the research process now for some of these adjustments and am very excited for what CAPS could end up turning up.

Finally, I'll occasionally mix in a few somewhat subjective feelings based upon PITCHf/x data. If I really like the pitches a guy is throwing or something else the about the data, I might put a little asterisk next to his name on my sheet, even if he hasn't totally proven himself yet. Call him a sleeper or an upside pick or whatever you wish, but that's the general idea. Max Scherzer was a guy I really liked this year despite his lack of a real track record. I particularly loved the change-up he throws. Zack Greinke was another - four good pitches and a fantastic slider. Until we can really build some predictive models out of PITCHf/x, though, this kind of thing will be merely supplementary to the hard numbers, at least for me.