JA Happ and Why Velocity Matters
J.A. Happ’s Fangraphs page does not make for encouraging reading. A fly-balling left-hander sounds like a good fit for the park,* but ideally the M’s would want a bit more than a generic label in exchange for a good cost-controlled OF, whatever his injury history. Happ’s home parks have hurt, no doubt, but it’s the combination of high walk rate and high home run rate that have really made his career FIP a mess. Unlike, say, Chris Young, Happ doesn’t have a history of beating the fielding-independent stats – he did, rather famously, back in 2009, but that looked like a fluke, and at this point, I think it’s pretty safe to call it one.
There is something worth talking about here – a reason for hope beyond the generic “lefty in Safeco” tag. Happ’s throwing a lot harder than he used to. This is something I talked about when he faced the M’s last year, and it’s something Jeff’s mentioned in his analysis of the deal over at Fangraphs. When he debuted with the Phillies, Happ’s fastball averaged around 89-90. Last year, it was around 93, well above average among qualified starters. Just to be clear, Happ is 32 (he played last year at age 31). This isn’t supposed to happen, but it keeps happening – Brandon McCarthy threw 89-90 in 2008, then 91 with Oakland, and last year, at the age of 30-31, started sitting at 93, and touched 95 occasionally.
But so what? Happ used to throw 89 and was bad. Last year, he threw 93 and was still pretty bad. Is this another case of people overrating velocity? Well, it matters because hitters, as a group, fare much worse against fastballs thrown faster than 92. This isn’t earth-shattering research or anything. But it’s not just whiff rates – batters slugging percentage drops when velocity increases.** Here’s a table of league-wide slugging percentages off of hard pitched (four- and two-seam fastballs, plus cutters) both faster than 92 and slower than 92 (data from BaseballSavant):
|League Ave||SLG% on FB> 92||SLG% on FB<92|
That’s an average gap of about 65 points of slugging, and as you can see, the trend is downward, especially for the faster fastballs. Despite league-wide velocity rising a bit, hitters are still having more trouble with plus-velo, or what used to be plus-velo and is now a shade above average velo.
Ah, but that’s cheating, you say. By slicing it that way, you add in all of the high-octane relievers and elite power arms like Strasburg, Fernandez, Richards and Harvey. Let’s look at some pitchers whose fastball averages around 92 and see what THEY look like with the same criteria – slugging percentage on fastballs above and below that 92mph mark:
|Player||SLG% on FB> 92||SLG% on FB<92|
Same thing. The range is a lot higher, but the pattern is very consistent (except for Henderson Alvarez, who remains baffling to me). Happ and McCarthy fit the pattern, though obviously the sample size differences are huge (they just recently became capable of throwing 92). Felix is awesome in every context, which is why we love him. But look at Sonny Gray and Phil Hughes! Elias’ splits are hilarious, but again, the sample is tiny. Or look at Kershaw, whose splits here mirror the league-wide numbers, albeit shifted lower. This is not a hard-and-fast rule, but maintaining good FB velo should help Happ.
So why didn’t it help him last year? In part, it’s because he had some bad luck on his other pitches. In part, it’s because he pitched in the AL East, home to a number of good hitters’ parks (his road stats were worse than his home line). In part, it’s because, despite the velocity, he’s not a great pitcher. Still, given the overall numbers, you can see why the M’s might see Happ as a good fit. The value of the pick-up is still, uh, debatable given salary, control and Saunders, but Happ may be better than he’s looked.
* So, about those park differences. You all know that Toronto inflates HRs to LF while Seattle suppresses them, but Tony Blengino’s granular batted ball park factors – something we get glimpses of in his articles at FG, are still something to behold. In this piece on Michael Cuddyer, Blengino includes a table of the park factors for fly balls to left center field. Toronto inflates production on such fly balls by just shy of 30%, as fly balls produced about 130% more runs than expected, given velocity and angle. Safeco? Safeco annihilates such balls in play, as actual production was just *36%* of expected given their launch angle and initial speed. 36%! To be fair, Happ’s HRs came more down the line than in the alley, but Safeco *also* suppresses doubles, which Happ also struggled with.
** This makes some sense, but may be counterintuitive to those who grew up on people saying the “pitcher supplies the power” and watching replays of Mark McGwire’s 500’+ HR off of Randy Johnson.