May 27, 2005
Lies, Damned Lies
Two seasons ago, I established a method to revise PECOTA projections on the fly by combining a player's PECOTA percentile forecasts with a binomial distribution to come up with a revised estimate of a player's "true" level of performance. The method was described in a lot of detail in the article that I've linked and I'm not going to describe it again here (plus it's late and I'm tired: do you know what your favorite baseball columnist is doing at 5 in the morning?). Long story short: this method is designed such that we have an effective way of gauging just how much an unexpected early-season performance should have on our assumptions about a player's long-term value.
With that in mind, we'll look at five players who have radically outperformed reasonable expectations, and five others who have radically underperformed them.
Current EqA: .375 (207 PA)
PECOTA didn't see this coming; whether or not anyone else did is another matter. In fact, the system saw Roberts as kind of a low-risk/low-reward kind of guy. There were some things to like about Roberts--prior to this season, he did essentially everything well except hit home runs--but nothing to indicate that something like this was possible. I know that some people have cited the 50 doubles that Roberts hit last year as post-facto evidence of some sort of impending power surge, but as Rob Neyer has pointed out, there really isn't anything magical about doubles that predicts a big jump in home runs. Yes, a certain percentage of those doubles "turn into" home runs--there is some predictive power there--but others turn into singles, or outs.
But given that Roberts did do everything well before he started hitting home runs, and that now he's hitting home runs too, I think we're looking at a guy who is probably among the 15 or 20 best players in baseball going forward. This method pegs him at a "true" EqA of .285, which would make him about a Ray Durham kind of player, but I think that's too conservative: as this graph makes clear, Roberts is absolutely obliterating even his 95th percentile projection:
The precedent that I think works reasonably well here is Jim Edmonds, who did not hit for much power until he was 25 or so, but did a lot of other things pretty well, and didn't look back once he did start hitting the longball.
Current EqA: .306 (190 PA)
Inge's season hasn't received an especially large amount of attention, but I chose him for this piece since I was skeptical of his 2004 performance, at least in private conversation. PECOTA wasn't especially optimistic either, although it recognized the weirdness in his performance record and marked him as a high volatility guy with both a high breakout rate and a high risk of collapse. There is an obvious exogenous factor here, which is the position change; the effects of position switches on offensive performance is a study that really needs to be done. I suspect that it also matters that Inge not only is playing a less difficult position, but is also playing it pretty well.
Current EqA: .354 (184 PA)
Lee had a lot of breakout comparables in his profile, mostly a result of his having stayed very healthy and having excellent secondary skills. The walks are back this year; whether that's the doing of new hitting coach Gene Clines or happening completely in spite of him I'm not certain. I had Lee pegged as a darkhorse MVP candidate prior to the season and that's looking fairly prescient, at least if the Cubs can somehow stumble their way back into contention.
Current EqA: .317 (178 PA)
PECOTA pretty much nailed this one. In fact, it doesn't think that Burrell's 2005 performance has been too out of line at all:
PECOTA often predicts breakouts for players that have struggled with low batting averages but have their other skills intact, as it recognizes that BA can be fluky. What we're seeing here is essentially the complete reverse of Burrell's awful 2003 season:
BA OBP SLG 2003 .209 .309 .404 2005 .311 .416 .536
Notice that if Burrell had somehow hit .311 in 2003, the OBP and SLG numbers would come out about the same as they've been this year, perhaps after making a slight adjustment upward in the slugging rate as a result of his move to Citizens Bank Park.
Current EqA: .345 (162 PA)
There are a couple of players who have had really uncanny walk rate spikes this year; David Dellucci is the obvious one, but Overbay works here too. Although Overbay is seeing a lot more pitches this year--4.11 per plate appearance versus a career average of about 3.75--I have some concerns about just how sustainable that OBP jump is. Overbay is a good hitter, but he's not so dangerous that you need to pitch around him at all costs, and I think some pitchers are going to counter-adjust by challenging him a bit more once he makes his second and third trips around the league. If that happens, he'll continue to hit for plenty of power, but the walk rate will drop significantly. I expect Overbay to finish the year at about .305/.420/.540.
Now, for the decliners:
Current EqA: .217 (189 PA)
Chavez' performance has been every bit as surprising as Brian Roberts':
I'm completely befuddled here. Chavez was my MVP pick prior to the start of the season, and he has no obviously exploitable problems in his profile, except for his persistent problems against lefties (Chavez isn't hitting lefties this year, but he isn't hitting righties either). The base hits at least have come back during the past ten days or so and it's tempting to wonder whether there was some kind of undiagnosed injury problem at the start of the season. This exercise is not kind to Chavez, lopping a full 25 points off his long-term EqA projection, but I'm going to go on record as saying that I think he deserves the benefit of the doubt, and will be fine long-term.
Current EqA: .221 (112 PA)
We do have an injury explanation here; Thome was bothered by a bad back for all of April and hasn't played at all in May. Yet in spite of that, and in spite of the fact that we aren't working with as large a sample size here as for the other players, my intuition is that he's lost a lot of his value for good. For one thing, back problems are about the worst thing that can happen to a hitter, the equivalent of elbow problems for pitchers. For another thing, Thome is obviously not much of an athlete, and things like his batting average and defense had already begun a pattern of decline. Players with this profile tend to lose it all at once--look at Mark McGwire in 2001 or Boog Powell in 1976 or Frank Howard in 1972.
Current EqA: .193 (155 PA)
Catchers are erratic performers, sometimes developing late, sometimes collapsing early, and generally going every which way but straight. There are a lot of catchers with Matt Nokes and Rick Wilkins types of track records and though Martinez had a better minor league history than either of those guys, I have some real worries about a slow-footed catcher who had derived a fair amount of value from his batting average. PECOTA was skeptical too--note the high collapse rate. I think the Indians seriously need to consider a position change, and I don't think it's a coincidence that Martinez--like Nokes--doesn't field his position especially well.
Current EqA: .215 (200 PA)
Aaron Schatz asked me about Furcal when I talked to him last night; to be honest, I hadn't realized quite how poor a season Raffy was having. Furcal has always profiled oddly, sometimes taking walks and sometimes not, sometimes appearing to be on the verge of a power spike and at other times regressing. The stolen bases have been there this year and his defense has rated well, so I don't think this is a case of a player's body suddenly giving out on him completely, as can sometimes happen with speed-oriented middle infielders (it's this phenomenon that was mostly accountable for the high collapse rate that PECOTA assigned to Furcal). I'm drawing a connection here that might or might not be fair, but I see the erratic performances, the pair of DUI arrests, and I wonder if everything is okay with this player.
Current EqA: .219 (178 PA)
Some preliminary work that I've done suggests that flukish performance in contract years is a real phenomenon. Statheads might be guilty at times of treating ballplayers as automatons but economic theory suggests that players really do have every incentive to post the very best numbers they can when free agency is impending, especially in a world where one good season nets a player a $64 million deal that is widely lauded as a bargain.
Whatever the reasons for Beltre's performance, it's virtually impossible that a player with a player with a "true" EqA on the order of the .332 that Beltre posted last year could have a run of performance this bad, even over just seven weeks or so:
Beltre's revised EqA estimate is at .263, which is almost exactly at the baseline he'd established between 2001 and 2003, when his growth appeared to have stagnated. I see absolutely no reason to take anything other than the Occam's Razor explanation: Beltre's 2004 was one of the biggest one-year flukes in baseball history.