The Occam’s Razor of Events

In Major League Baseball, Dan Uggla of the Florida Marlins is a Most Valuable Player candidate this season. In 2005, he was a non-prospect for the Arizona Diamondbacks. The Marlins didn’t have a real option at second base the following year, and acquired Uggla for almost literally nothing from the Diamondbacks in the Rule 5 Draft. No one thought much of it, as he was considered a marginal player who posted only an ostensibly impressive performance in the minor leagues despite his advanced age. Sabermetrics think tank Baseball Prospectus published in their 2006 annual that

While on the surface he might look like a hustling, dirtier version of Tony Graffanino, there are a few cautions. He spent three years in Lancaster`s bandbox, and only “broke out” after an extended stay at Double-A. He`ll also be 26 by Opening Day. His history of “eventually getting it” made him one of Florida`s Rule 5 picks. With the Marlins he has a chance to start at second. The bar is low, but considering he`ll move from low minors hitter`s parks in offensive leagues to a major-league pitcher`s park, the results should be predictably Uggla.

Tony Graffanino was an occasionally useful second baseman who spent much of his career as a utility player or super-sub off the bench. His best years were around average defensively and offensively for the position. This was not a flattering comparison.

Uggla then went on to finish third in the league in Rookie of the Year balloting, followed with a year of similar value in 2007, and has had his best year so far in 2008.

Why did Uggla’s performance jump so far ahead of expectations?

Occam’s Razor states,

that the explanation of any phenomenon should make as few assumptions as possible, eliminating those that make no difference in the observable predictions of the explanatory hypothesis or theory.

Practically, it is taken to mean that the simplest explanation is the best one. Instead of strong, convoluted theories, we should understand and start with the basic one. There are millions of hypothetical explanations to any set of circumstances, so we have look somewhere. Although the principle is mostly applied to science, I am interested in its application to history, the true causes of events, and to avoiding Monday Morning Quarterback as much as possible.

The simplest reason why anything happens is no reason. No underlying fact needs to change for reality to be different. What we see now in the living world today may often be pure noise and no signal.
Angel Berroa is another rookie who performed well above expectations. In a hotly debated ballot, he won the 2003 American League Rookie of the Year. Before 2003, Baseball Prospectus said

Let’s see: the Royals’ best prospect entering the season injured his knee and missed two months, hit .215 in Omaha after he returned, showed vastly diminished range, and drew six walks by the end of July. On the plus side, he did celebrate three birthdays last year [due to crackdowns after 9/11, players “got older” when it was revealed that they lied about their birth certificates. It is strongly believed younger players have potential to increase the skills they already have. I looked but cannot find a good citation for this, although you can find references to it by googling the catty neologism, “agegate”]. No, it was not a good year for Berroa. His apologists argue that he continued to favor his knee when he got back, which hurt him both offensively and defensively. In their support, he looked much more comfortable both at bat and in the field during a September call-up. The Royals are determined to let him take over as the starting shortstop this season, and given that he turns 25 in April, they might as well see if he’ll sink or swim. There’s upside here, but right now Billy Beane wouldn’t trade Mark Ellis for him straight up.

No one took Berroa that seriously prior to 2003. After that sparkling rookie season, he went on to play worse until he was demoted in favor of someone who, if nothing else, had reasonable defensive credentials, in Tony Pena, Jr.

The simplest explanation to this set of events is that Berroa was more or less the same player throughout the entire period. Nothing caused him to have above average skills across the board in his peak season except pure randomness. This explanation is far simpler than beginning to suggest that he “broke through” in 2003 and later “lost bat speed” or that “pitchers found holes in his swing.” Announcers exhibit the pretense of “knowing” why certain pitchers choose to throw a certain pitch or why a hitter couldn’t come through in a certain situation, while questions of numbers and averages smoothing out seem invalid over the course of even an entire season. The null hypothesis explaining Berroa’s 2003 should be that there is no reason at all.
If we have evidence that something actually changed and caused a result, it is more likely that a “Confluence of Circumstances” caused it rather than a single event. Although this may seem unintuitive and at odds with the previous presumption that normally nothing changes, it is in the same spirit. It’s far easier for a multitude of minute “real” changes to occur in each underlying variable than catastrophic change in one. For any one of these changes, a biased author may attempt to explain the entire event by pointing to one while then assuming the others. It may be true that the event wouldn’t have occurred without one of the causes, but that does not imply that the one is the only cause.

For the classic example, consider the Roman Empire. Quite a few people have already. Would Rome have fallen were it not for the Germans? Probably not. A continued reliance on slave labor and the Antonine Plague? Perhaps. It is most likely that not one of these causes by themselves, but some combination thereof, caused Rome to fall.

Whenever you hear of a city in riot or a failing company, it is safe to assume that it would not have happened had it not been for several other adverse conditions affecting it, rather than the one or two the reporter happens to mention.

People look for simple answers. Those answers aren’t the truly simplest empirically or logically, but the easiest to understand. Just as unintuitively, those “simple answers” are the absolute opposite of Occam’s Razor. It’s far easier for unpredictable, subtle movements in several variables in data or a slight alteration in several factors causes an event than some narrative event to push history forward.

There are certain counterexamples. When the Federal Reserve in the United States decides to raise interest rates, there’s not a lot we can do about it. There are myriad reasons why it may choose to do so, but when it does, all it needs to do is decrease the supply of money until we reach that point. While the question of why the Fed decided to raise interest rates remains, it’s clear enough, with certain theoretical exceptions, how it increased after the Fed made the decision.

The most unlikely scenario is the conspiracy theory. Not only did someone see whatever happened coming, but planned it and used it to his advantage. The strongest implication to make is that of evil. It requires the assumption that whatever happened did not occur due to randomness, that a single variable can control the result, and that someone is able to take advantage of it. Still, there are different “kinds” of evil. It’s reasonably believable that in a controlled setting, such as the votes of congress, that strings can be pulled by lobbyists and the decisions of the few can cause drastic changes for the many. In an uncontrolled setting, such as that of public opinion or the free market, conspicuous evil is next to impossible. When the public is not coerced physically or monetarily into rejecting conspiracy theories (read: the explanation that someone is doing something evil without fear of retribution), it’s more reasonable to point to the general un-truth of the world than, say, the price of oil is high because the companies are trying to screw you.
To summarize, Occam’s Razor slices several distinct layers.

1. It’s simpler to say that randomness caused something to occur than any set of events.

2. It’s simpler to say that a several events caused something than a single event.

3. It’s simpler to say that a single event caused something than that someone caused that event maliciously.

3. It’s simpler to say that someone caused it maliciously than to say that someone caused something maliciously and no one caught onto it while it was happening.

Where does Dan Uggla fit into all of this? Some may suggest he got discouraged by staying in the minors for long while many talented prospects in the Diamondbacks system shut down his future. Others may point to the workouts he undertook after finding out he was drafted by the Marlins or the quality of Major League Coaching he finally received. At this point, with the empirics we do have, I reject the assertion that his improvement may still be pure randomness, but it seems overwhelmingly likely that his radical increases in performance were caused by dozens of inscrutably slight adjustments in place of something that the casual baseball fan can comprehend in print.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


Get every new post delivered to your Inbox.

%d bloggers like this: