Last season, the question was, what's wrong with Pronk? This season, the question is, is there any reason to think he's coming back? And by "he" I mean, the old Pronk, the young one, the Old Young Pronk, the Scary Monster. Hell, you know exactly what I mean, why am I trying to explain it?
All the major projection systems are calling for a mild bounce-back for 2008, but this is cruelly neither the comfort (great chance he'll come mostly all the way back) nor the condemnation (he's toast) that we crave. I mean, guy has shocking off-year, he'll probably be somewhat better this season, but probably not as good as he used to be -- Duh. Like I needed your goddam algorithms to tell me that.
UPDATE: For more goddam algorithms, see also Adam's examination of Hafner's balls in the Diary section, an equally detailed, interesting and futile take on the subject. [Jay]
I've got my gut instincts about players like everybody, but I only bring them out into the sun once in a while. I just don't believe in arguing about our guts -- my gut is probably bigger than yours but not necessarily better. And I don't want to just swap guesses all day -- part of the day maybe, just not all day -- and I do want to understand. What's happening to these players? What's going to happen? And how much, and how likely, and why? So yeah, I do mostly talk about the objective stuff.
But for the love of God, how can we be objective about the Pronk? I mean, we can, but how can that satisfy us? The Quest for Pronk is a philosophical stake in our hearts, confounding our minds and flashing our collective ignorance, piled atop a trash-heap of statistics, in bright neon letters. What do we really know? NOTHING. <off> NOTHING. <off> NOTHING. <off> NOTHING.
The heart thing, I can't help anyone there. We're all in it together, and that's the only comfort we're likely to get. The brain thing, though, maybe I can be of some help. At the very least, we can take a look at what the projections do say and don't say, and what they can and can't tell us. As I mentioned earlier, every system will sheepishly offer up a mild bounce-back -- regression, you ain't got a system without it -- but in PECOTA, the multi-year trend line is definitely a decline, as I think it would be in any system considering more than one season forward. If all you knew about Hafner was what PECOTA knows, 2008 was a great year for a team to have his contract ending.
But let's step back from PECOTA and, possibly, off the ledge, just for a moment. Subjectively, I think you have to look at just how strange and rare a case Hafner presents. It's all well and good to throw up a couple of comps like Boog Powell -- I've done worse -- but good, robust projection involves 100 or more somewhat similar players, because we're trying to peg what types of player Hafner might be. One or two players do not a typology make, and realistically, Hafner '08 does not have 100 comps.
How many players have ever racked up three straight seasons with .305+ average, .408+ OBP, .583+ slugging, 563+ PA, 162+ OPS+, and 281+ total bases -- reaching every one of those statistical levels, in every one of those three consecutive seasons? I would guess that there have only been a dozen or so players who've done something like that in the past 20 years (and yes, four of them were Indians). It's not unheard-of, but we're talking about a handful of players.
Next, we have his 2007 performance, which (regardless of current projections) was completely unexpected. I don't have the old 2007 projections, but my recollection is that his 2007 line was well below the 10th percentile for his aggregated/weighted comps. I'll call it 5°, though it could easily have been 2°. (Maybe I'll ask Nate Silver.)
The point being, here you have a player who is one-percent rare to begin with, having a five-percent result season after that, and the net is one really baffling situation that can't be predicted with much confidence. It ain't like Konerko or Morneau, Grady or Grandy. It's a very, very rare case, and you can't tell me that there is a typology for Scary Monster Without Relent For Three Years And Then Bupkis, because there ain't.
As Ryan noted in another thread this morning, we're not looking at an "old player" skill-set here. Hafner is in the 90th percentile of all hitters based on batting average, so obviously he can't be considered a "low average guy," even if James really meant "average average." This only works if you think you can map typology based on one season, ignoring all previous, and obviously nobody serious thinks that. Hafner also shouldn't be considered slow, certainly not old-player slow. His being a DH has nothing to do with his mobility, as he goes first-to-third quite effectively -- yet another factor that makes him hard to typologize neatly.
Thing is, in the relentless pursuit of comps, there's no doubt Hafner has a bunch of old-player-skills guys mixed into his PECOTA projection, because they have to compare him with somebody. And even outside of total old-player-skills guys, he's mixed in with slow guys, gimpy guys. He's not slow and he's not gimpy, but again, the system just has to find some comps.
By the way, his being a very rare case is pretty scant consolation, because it doesn't mean he'll bounce back. Here's another weird thing about Hafner: for 2004-2006, his BABIP was trending down (.350, .344, .323) while his OPS+ was trending up (162, 168, 179) -- so you might say that his luck-normalized production was trending up even more dramatically. In 2007, his BABIP drops to .294, his OPS+ drops to 118. Chicken? Egg? Caviar? Tofu? Could he have been exceptionally lucky for three years running? Let's face it, that is a possibility, because it's a given that we're looking at a very rare circumstance, so possibly some very rare variable element is the main cause.
So where does understanding the rarity fit in with our processing of PECOTA? Well, objectively, PECOTA is what we've got. Beyond our guts, and beyond the front office's collective gut, we've got to deal with the range of probabilities that PECOTA lays out for us. That is the objective answer. The rarity comes in just to say: Either he is, or he ain't. Meaning, either he's never going to be Old Young Pronk again, or he is going to be the Old Young Pronk again and all of this was, more or less, hooey. N/A. Thank you, drive on through.
PECOTA actually agrees with this conceptually, if you dig deep enough. It traces very different career paths for Hafner at the 10°, 25°, 40°, 50°, 60°, 75° and 90° points:
- 10° and 25° -- a continuing rapid decline, he's finished by the end of 2009.
- 40° and 50° -- no significant bounce-back, a sharp decline in 2009, and he's finished by the end of 2011. (The 50° numbers are commonly used as shorthand for "his projection.")
- 60° -- a decent bounce-back followed by mild decline. He's not coming back, but he's not toast either.
- 75° -- a great bounce-back, reasonably regressing for age in 2008 based on his pre-2007 production, and continuing as such. Still a star hitter through 2011 and solid through 2013. This is what we expected of him a year ago.
- 90° -- Scary Monster is back, without even any age regression. He's at 2005 levels for 2008 and 2009, then settles into a typical slow decline, remaining a star hitter clear through 2013.
In other words, what PECOTA is really saying is, there's a 25 percent chance that he'll get back to being the player we thought he'd be or better, along with the 60 percent chance that we really have lost the Old Pronk, substantially, for good. Cold comfort, I know, but isn't it nice to know the Old Young Pronk is in the mix there, and not just infinitesimally?
If there's an up-note to end on here, it's the front office that signed Pronk to that suddenly-humongous-looking four-year, $57 million extension last season, even with his relative slump in full, mediocre swing. Every team knows more about its players than its fans do, because of scouting and medical information. The Indians are one of the few teams that also knows more than PECOTA does about the objective data. The Indians track hard performance numbers that PECOTA doesn't incorporate, and their internal projection systems are every bit as sophisticated and probably moreso. Diamondview was probably at least PECOTA's equal in this area two years ago -- and then the Indians hired Keith Woolner.
So while there are few front offices worth betting on ahead of PECOTA, the Indians are probably one of them. And that's basically all I've got for you to pin your hopes on: One very smart and sophisticated front office, and one totally livable 75th percentile projection from PECOTA. It ain't much, but it might be just enough for us to keep our sanity through May.
Hat-tip to one of the Nicks and to Chuck, for forcing the closer look.