Pitching staff moves, philosophies
According to Larry LaRue at the News Tribune, the M’s are indeed sending Ian Snell to the bullpen, sending Jesus Colome to Tacoma, and freeing up a spot for Cliff Lee in the M’s rotation. That makes sense given the early season – Snell’s struggled, while Jason Vargas and Doug Fister have pitched well.
The decision doesn’t surprise me in light of the results so far – both Vargas and Fister have been surprising to date. They run together for me as fringy rotation starters who saw significant time with Tacoma in 2009. Both are primarily FB/Change-up pitchers, with fastballs in the 87-89 range (Fister’s got a touch more velocity, but Vargas is a lefty). Through four starts, both have made big changes to their pitch mix. With samples this small, it may not mean anything, but it’s interesting to ponder.
Doug Fister’s fastball got poor results in 2009 according to Fangraphs‘ pitch type linear weights. This shouldn’t really surprise anyone, given that he’s a righty with an 88 MPH ‘heater.’ He generated well below-average swinging strike rates in AAA and MLB last year as a result. His change-up, however, generated more swinging strikes and better results. So what’s changed in 2010? To date, he’s throwing a lot more… fastballs. And it’s working! Intentionally giving up swinging strikes in favor of batted balls isn’t generally a smart play, but with this defense, it’s *almost* understandable. He appears to be trading swinging strikes/K rate for fewer walks. The cost of this shift is an awful lot of balls in play, which, with this defense, sort of makes sense.
The results look odd: Doug Fister’s FB fools exactly no one. His sub 2% whiff rate with the pitch is almost unheard of, as is his FB frequency: a league-leading 80%. So far, so good – his results on the pitch are quite good, though the pitch results with a sample like this essentially act as a proxy for his BABIP, which is an absurdly low .241. More encouragingly, he’s getting a few more swings outside the strike zone (O-swing%), which typically result in poor contact (the best result you can hope for if you’ve given up trying to generate swinging strikes). That’s all well and good, but a glance at pitchers with sub 5% swinging strike rates doesn’t inspire confidence. Essentially, you’ve got high-GB pitchers like Joel Pineiro (who was still above 5% last year), and guys like John Lannan and Livan Hernandez. These days, if Livan Hernadez shows up as a comparable pitcher, you’re doing something wrong. This approach doesn’t make sense, but it IS familiar.
Last year, Sean White threw FBs about 75% of the time (for reference, his FB% in 2007 was under 60%), and got similar results: 3.7% swinging strikes on a pitch he threw almost exclusively. So far in 2010, he’s purged a few more swinging strikes with his FB. David Aardsma found success for the first time in his career by increasing his FB percentage significantly last year, and so far in 2010, he’s essentially trying to make people forget he even has offspeed pitches. This is interesting given that these three pitchers don’t resemble one another in any way. If this is something Rick Adair preaches, what’s the idea? White and Aardsma throw hard, but Fister doesn’t. White and Aardsma have battled some control issues, but Fister was always a control artist. White was a ground-ball guy, while Fister and Aardsma have yielded more fly-balls.
So what about Vargas? The other FB/Change-up pitcher with no velocity had decent results on his change last year, and, pace Fister, has increased his change-up frequency by nearly 10 percentage points in 2010. He’s also generated more swinging strikes and a higher O-swing percentage. Why the stark differences in approach by Fister and Vargas? Some may point to their catchers; Fister’s FB-dominance has coincided with his last three starts with Rob Johnson catching, while Vargas has had Adam Moore behind the plate for his starts. Another possibility is that the M’s think certain pitches (technically, certain FBs) have lower average HR/FB rates, and essentially punt Ks to get more balls in play. If you don’t walk anyone (like White and Fister) and don’t yield HRs, this makes sense, but it runs counter to sabermetric wisdom that HR/FB rates regress towards the mean. That’d be a fascinating result, given not only the amount of data we have on HR/FB variability but also Fister’s so-so HR/FB rates in 2009. That said, I’m not aware of many studies on HR/FB by pitch type – and our current pitch types may be too broad in any event (See this interesting study on pitch-type platoon splits by Max Marchi).
Does anybody have any alternate theories of why Vargas and Fister’s approach would diverge like this? Why would the M’s look at Fister’s 2009 results by pitch and instruct him to throw fewer change-ups? Why would they do so while telling Vargas to throw *more* change-ups? Vargas has faced more opposite-handed hitters, so that’s probably a factor, but one noticeable change from last year is Varas’ willingness to use his change (and not his slider) to lefties. He threw change ups to lefties fewer than 4% of the time last year, and now it’s about 15%. What do you think is going on here? Are the M’s actually harming Fister’s chances of sticking in a major league rotation with this pitch mix strategy, or have the M’s using a novel approach that might make Fister serviceable at this level? Given the sample sizes we’re looking at, there are no definitive answers – but what do you think’s going on here?
(Edited to add Vargas change-up % by handedness and correct weird error about throwing change-ups to same-handed hitters. It was late.)