Breakdowns

I love baseball for its never-ending parade of surprises – and the community surrounding baseball is no different.

We had easily our biggest day of traffic yet on Wednesday, thanks to Drew Fairservice, Tom Tango and the Blue Jay Hunter. Thanks given, hats tipped, beers owed. The driver of the traffic was my (surprisingly contentious) resistance to the new Shutdown and Meltdown stats introduced by Fangraphs. So, in the spirit of Science!, I’ll give it a shot on its own terms. I’ve even got a good test case lined up:

Jason Frasor.

(You may wish to load the Fangraphs page linked above in another tab to follow along.)

Frasor is something of a whipping boy for Jays fans (and that’s putting it mildly, if Mike Wilner’s Jays Talk is any indication). He’s a been damn good pitcher the last couple years, by most measures – he posted ERAs of 2.50 and 3.68 in ’09 and ’10, respectively, with an FIP of 2.99 and 3.31, FIP- of 68(!) and 80, and xFIP- of 82 and 89. (Side note – I rather adore the new minus stats Fangraphs introduced. You should too.) He has a career K/9 of 8.42, and a K/BB of 2.21. He’s somewhere between “very good” and “excellent”. There are good reasons he was a Type A free agent this past off-season.

And as I said last post and Stoeten mentioned post-game, he’s excruciating to watch and not trusted at all by the fans; Mike Wilner defended Frasor during his recap, struggling against the tide of commenters and Twitterfolk with their visceral hatred of the pitcher. So, let’s bring out some stats, including the SDs/MDs I badmouthed last post, and see what we can see.

Jason Frasor, 2009-2010:

Hrm.

Frasor was excellent in 2009. ’09 was also the premature end of Uncle BJ’s Wild Ride, so Frasor picked up a number of the save opportunities vacated by the former closer. He did much more to win games than his saves and holds would attest, and as the cumulative WPA total (2.62(!)) indicates, Good Things Happened when he was on the mound. Last season was a little rougher on him. His FIP nudged up, but was still within the “very good” category – the difference was in the results. Note the drop in SD/MD, and especially note the cumulative WPA of -0.13. Negative overall WPA. Whatever things Jason Frasor was doing on the mound (mostly walking around in circles and sighing heavily, it seemed) he didn’t manage to translate into increasing his team’s chances to win. (I’ll let you go through his Fangraphs page and draw your own conclusions as to why – I’ve got mine, but since they’re only suspicions, I’ll keep them to myself for now.)

Net result: scapegoat, whipping boy, lightning rod.

Let’s stack his performance last year against some other members of the Blue Jays bullpen:

Conclusions, or at least fumbling attempts at such:

- Poor Casey. So many innings, so many ways things could (and did) go wrong, so little chance to contribute. (That was Cito for you.)

- Note how many empty saves Gregg got, how much damage he did overall, and thus how justified our scorn was. (And by god, there was scorn aplenty.)

- I liked Camp last year, despite the (literally) middling FIP. He was master of the high-leverage ground ball to get the double play, and the WPA reflects that.

- Scott Downs was awesome at putting out fires.

- Jason Frasor, despite being a good pitcher and having roughly the same SD/MD ratio as Downs and Camp, just didn’t have enough Good Things Happen when he was on the mound. And that sinking, gut feeling like Frasor’s appearance would leave things worse off? Somewhat justified, on the whole – certainly compared to his superlative 2009.

And the point of the exercise:

Four pitchers with a whole bunch of innings and a substantial spread in FIP and cumulative WPA also all have SD/MD ratios within spitting distance of each other. I’d hazard a guess that somewhere around 2:1 SD:MD is pretty common for relievers in the ~60 IP range, at least if they pitch innings that matter (Janssen’s 2010 pLI was a mere 0.44 – mop-up and filler appearances, pretty much exclusively). Someone would need to assemble much more data than I have, of course.

What I do see, however: cumulative WPA probably has a pretty good correlation with general fan perception of a relief pitcher, seemingly more so than shutdowns or meltdowns or even the ratio between the two. And not all meltdowns are made equal in the Big Picture; since the minimum WPA change between entry and exit is a mere 0.060, doing a little poorly is tallied the same as outright losing the game entirely. Looking at the overall WPA for a season is probably a better bet if you’re trying to gauge fan favorites, and FIP- is a sweet little tool for measuring skill.

What I don’t see, still, is the usefulness of shutdowns, meltdowns, or comparing the two. At least, not without looking at a bunch of other stats at the same time. And maybe not even then. I think the cutoff is perhaps a little too low. I think it tells me nothing other stats don’t, and might even be a little misleading. I think there’s a gap between what a pitcher should do in a situation and what happens when he does it – which should be explored further, and which shutdowns and meltdowns fail to do. I think WPA in general could benefit from more granularity to illuminate the separation of process and results, of skill and success. (I’ve got a couple ideas – I might post them at some point, once they’ve rattled around my head a little longer.) What I won’t do from here on, though, is bitch about it further. Consider me a skeptic and leave it at that.

And don’t ask me what I think about Jason Frasor. I’m more confused about him now than when I started.