February 28, 2012
Random processes produce many sequences that convince people that the process is not random after all. —Daniel Kahneman
With the Oakland A's struggling recently in the wake of Moneyball and its chronicling of their success despite the odds against them, a narrative suggesting that Billy Beane's strategies no longer work has evolved. The theory is that with Beane's secrets exposed to the world, he has lost the element of surprise in attempting to defeat the big boys and their big pocketbooks. Other teams have become savvier, and Beane cannot exploit market inefficiencies the way he once did because there are fewer of them.
It's a compelling narrative, and one that doubtless contains some truth, but there is another possibility: regression to the mean. In succeeding against teams that had greater resources, the A's were an outlier. This is why a book was written about them. Nobody wants to hear about a fully average experience. That story doesn't sell.
Without wishing to denigrate Michael Lewis' fine book, Moneyball was, at its core, a collection of entertaining tales that maybe made folks without advantage feel better about their chances in what can be a difficult world. It also introduced many people to sabermetrics, which is wonderful. But the tales wouldn't have been so compelling had the A's not enjoyed an unexpected run of success. These were extraordinary events, which means that, by definition, they were not ordinary.
In general, we would not expect teams that lack the wherewithal to procure the game's best players to compete with those that can afford top talent. And in general, poorer teams don't compete with richer teams. But occasionally, Beane's A's happen. Or some various incarnations of the Twins or Indians.
These teams and the folks that built them deserve credit for their success, of course, but any success is a combination of good work and good fortune. Luck eventually runs its course, and teams that are able to rise above their limitations for a while return to a more humble existence. (Some teams, e.g., the Pirates and the Royals, don't even get that much.)
As it is possible to construct a narrative that explains the A's rise to prominence, so is it possible to do the same for their fall. Based on what we know now (e.g., everyone is hiring a sabermetrician or three these days), we can make Oakland's demise appear inevitable. We can begin sentences with, “As should have been obvious...” or “Although it didn't seem important at the time...” and make ourselves look smarter than we are. Reconstructing the past is a lot easier than predicting the future.
Then again, predicting the future (or attempting to do so) can be fun. As long as we recognize its limitations and are honest in later assessments of our performance, there is no harm. With this in mind, I offer three bold predictions for this year's A's. I don't necessarily believe that all (or any) of these will happen, but as they say, go big or go home. If I hit on one, I can claim to be brilliant and ignore the other misses—that's a good analyst trick.
It's also crap, so we'll take the road less traveled and go for accountability instead. After the season, we'll review these predictions and, well, maybe you can have a good chuckle at my expense. Then I will shirk all responsibility by reminding you that the future is inherently unknowable. It'll be great fun.
But enough of my yammering; let's get to the good stuff. Here are three things that probably won't happen, but conceivably could happen, in Oakland in 2012.
1. The A's Will Win Exactly 75 Games
Aside from 2010, when they finished at .500, the A's have been consistent. Still, predicting exactly 75 wins is crazy stupid. It's similar to picking 22 at roulette (there are 38 possibilities on the wheel, while the average spread over the past five seasons in the AL has been 36.4). The probability of any single win total occurring is remote. You're better off going with a range, like 74 to 81 (or, if you want to play it real safe, 64 to 91, which isn't too far off last year's AL range), but where is the fun in that?
So we'll go with the single point. Why 75? Why not? We have to pick something; it might as well be a number we know the A's are comfortable with. It might be more fun to go with 91, but that seems too far beyond the realm of reasonable possibility. We could get 75 wins and look smart. And we could find “supporting evidence” for this prediction: “Well, if you figure that Kurt Suzuki will... and a full season of Jemile Weeks... plus Cliff Pennington is still young... but pitching is a concern after the trades this past winter... and the Angels and Rangers will be tough.”
Add enough detail (as Kahneman notes, “a compelling narrative fosters an illusion of inevitability”) and voila, 75 wins. Or whatever number within a tight range floats your proverbial boat. It's plausible. And if we do it right, we have excuses for when the A's win some other number of games. “Well, Suzuki didn't... and Weeks didn't...” or, if we predicted too few, “The rotation looked iffy in the spring, but who could have foreseen the emergence of Tyson Ross?” Whatever the outcome turns out to be, adjust the original narrative slightly to accommodate events that occurred after the fact. Genius!
2. Tyson Ross Will Have a Breakout Season
PECOTA also gives Ross a 23 percent chance of breaking out. This isn't the highest number among A's pitchers (A.J. Cole is at 50 percent, but let's wait until he's mastered Double-A), and Ross isn't the sexiest breakout pick (newcomer Jarrod Parker is at 33 percent), but the potential exists. Besides, we have to pick someone, and I'd just as soon let it be someone that everyone else isn't picking. (This is what differentiates our breakout from that of others. The Ross pick is slightly—how shall I say—insane.)
Ross' comps are mostly uninspiring. Wade Davis? Yawn. Frankie De La Cruz? Not even a yawn. But wait, here's a fun name: Mike Witt. He wasn't great, but that dude could pitch a little. Witt made 299 starts over parts of 12 seasons (mostly with the Angels) and was an absolute beast in 1986, at age 25. Oh, hey, did I mention that 2012 is Ross' age-25 season?
This is a terrible argument. I'm making a point about how to construct a narrative, but you should be aware of a key difference between Ross and Witt coming into their age-25 campaigns. Here is how they had fared to that point in their careers:
So, yeah. For comparison, would you like to know what I was doing at age 25? Because it probably makes as much difference to Ross' future as what Witt was doing, but never mind that. We've found our narrative and we are making bold predictions.
First prediction: Ross will work more big-league innings than De La Cruz did at age 25 (3.1) but fewer than Witt (269). This isn't the bold part... start slowly and build from there.
Second prediction: Eh, let's just go for it. Here is Ross' CRABS (CRazy Ass Breakout Scenario) for 2012 as compared to his PECOTA:
CRABS won't happen, but if it does, you heard it here first. Because who else would say such a thing?
3. Manny Won't Be Manny
In the dream, I was standing on the cliffs at Big Sur, not far from Esalen. A man approached me and asked if I wanted to expand my vision, to know the truth.
He was surrounded by Monarch butterflies. I said yes.
The man told me that everything I needed to know was contained in this simple formula:
Pns = (a + b) * c
He explained that Pns, the probability that nothing stupid will happen, is based on three factors. I was dreaming when I wrote this; forgive me if it goes astray:
The original exponent was 2, although rigorous testing has shown that 1.83 is slightly more accurate. Ordinarily, we would use the previous season's statistics, but since Ramirez took a “vacation” after his five-game stint with the Rays last year, we'll go back to 2010:
a = (42 – 6)1.83 + 6
This, then, is the first part of our equation. I don't understand why these items should be related at all, let alone in this particular way, but remember: the man was surrounded by butterflies (the devil, or a compelling narrative at least, is in the details).
The b component is similarly baffling:
= (38 + 3 – 1)/16
So we now have the first two parts of our equation, which gives us the number 713.2 (being the sum of 710.7 and 2.5). This is a very large number, indeed, more so than 500 but not yet as much as 1,000.
The genius of our equation lies in its final component, c. This is simply an adjustment coefficient based on a variety of factors that lay beyond my comprehension. Perhaps the man in my dream was too smart for me, or I was too dumb for him. Alcohol may have been involved. But I do know that the coefficient is a constant in the case of Ramirez, and that constant is 0.
So, completing our formula, we get the probability that Manny Ramirez won't do something stupid:
Pns = (710.7 + 2.5) * 0
In other words, there is no chance whatsoever that he won't do something stupid. And this is what makes my prediction so bold. I submit that despite overwhelming evidence presented to me in a dream by a man surrounded by Monarch butterflies, Manny Ramirez won't do anything stupid.
Also, he won't hit 44 home runs.