Real Improvements?

A couple weeks ago over at The Baseball Analysts, Dave Allen had a great post about how the speed of a pitcher's first 100 fastballs predicts whether they'll beat their FIP projection.  Since FB velocity normalizes so quickly -- the correlation between '08 and '09 FB velocity for the 85 guys I looked at is already 0.970 -- it's one of the more meaningful measures to look at early in the year, because it's less prone to small sample size flukes.  With that in mind I thought I'd take a look at which guys are throwing harder than they did last year, and how that's affected their FIP, as well as their walk and strikeout rates.

I took guys who threw at least 100 big league innings this year, and have faced at least 80 batters this season. Here are the five who've seen the biggest increase in their fastball velocity, along with the change in their FIP, K/9, and BB/9 (for each, a positive change is good; I flipped to sign for BB/9):

I haven't heard much talk about Zito's improvement, probably because the other guys on that staff are much more interesting, and an increase in velocity from 84.9mph to 86.7mph really isn't all that exciting.  It is significant though, and it's supported by accompanying boosts to his curve (70.9 to 73.7) and slider (77.7 to 79.3), although he's not throwing his change any harder.  His ERA was likely to come down from last year's 5.15 anyway, but if he keeps this up he could possibly even get back to being an average starter.

Verlander's stats have already been documented, and I would've looked a lot more prescient if I'd written this post before last night.  In his first four starts he wasn't getting results, thanks to a ridiculous BABiP (even after yesterday's dominant performance it's over .400).  His ERA still sits at 6.75; considering the increase in velocity (93.6 to 95.3), ridiculous K/9 (10.9) and high BABiP (.408), he definitely remains a great buy low candidate.

Galarraga is an interesting case.  He had a great rookie campaign last year, going 13-7 with a 3.73 ERA, but that looked to be rather fluky, with a 4.88 FIP and .247 BABiP.  He's off to a great start again this year, 3-0 with a 1.85 ERA, but this time there's actually some evidence that he may be able to keep it up, with his fastball going from 90.2 to 91.3, his slider up a tick (0.6 mph), and his FIP way down at 3.37.  His ERA won't stay under two of course, but there's reason to believe the previously inevitable regression back to league average (or worse) won't ever come.

Well it sure is good to see Billingsley on this list; that guy certainly needs all the help he can get.  For him, the improvement has translated (not that theres necessarily causation) into a lower walk rate, down to 3.1 from last year's 3.6.  He's also yet to allow a HR, which may not continue.  It's worth noting that he hasn't exactly faced offensive powerhouses thus far, going up against the Padres, Giants, Rockies, and Astros (and SF again Tuesday night).  That's another nice thing about looking at FB velocity; not only does it stabilize quickly, but it isn't biased based on level of competition.

Oswalt, with an average FB of 93.5 mph, is throwing harder than he has in any year since '04; he's also upped his curve from 70 to 73.5.  It hasn't showed in his results thus far, but considering he's never had a FIP over 3.80 in his career, that's probably not much to be concerned about.

I'll get to the guys who have seen a dip in velocity on Wednesday.  It's also interesting to look at how an increase in velocity affects different factors.  In terms of walks, it seems like there's not much of an effect; the correlation between FB velocity increase and decrease in walk rate is just 0.002 for these 85 guys.  It appears, at least in this data, that an increase in velocity goes straight to your K rate.  Here are the results of a regression with the independent variables of '08 K/9 and '09 FB velocity % increase, and the y-variable as '09 K/9:

Last year's strikeout rate tells most of the story, but that fastball velocity coefficient is enormous, if not quite statistically significant (yet). Now, we don't know that these guys will be able to keep up their velocity increases (I'm curious to see how much they can), but that means an increase in FB velocity from 91 to 92mph would be expected to be accompanied by an increase in K/9 of 0.39. We'll have to wait and see how much of this is sustainable going forward.

All stats courtesy of FanGraphs.