Hardball Times: When the wisdom of the crowds is wrong

Philosophers have been debating free will vs. predetermination for centuries. Many who subscribe to the latter point of view believe that the actions of an individual are the natural outcome of a series of fixed, historical events. In other words, free will is merely an illusion.

Anyone who participated in a fantasy baseball draft or auction before the season started may find that viewpoint surprising. After all, don't we exercise choice when determining things like whether to ignore Alex Rodriguez's injury and take him anyway in drafts or whether to take Nick Markakis or Matt Kemp in the 3rd round?

Maybe so.

But consider everything that seemed almost conventional thinking by the time we all showed up to draft. For example, Chris Davis was drafted anywhere between the 54th pick and the 80th pick in almost every draft, according to average draft data. Exactly who determined that Davis was this kind of value -- and how?

Understanding the answer to this question helps us make decisions when subsequent, unexpected things happen.

Last October, the 2008 baseball season ended. At the time, most people who participated in fantasy baseball turned their attention to other things -- like remembering their girlfriend's name. Or football.

Meanwhile, a group of hard-core enthusiasts started to assess the previous season. Many began conducting mock drafts. Pretty soon, people like Sean Smith and Bill James and organizations like Baseball Prospectus and BaseballHQ began considering things like a player's past performance, a player's age, and peripheral similarities to others who have played the game in the past century, releasing projections for what we might expect in the 2009 season. (Methods vary.) More mock drafts. Soon, the big fantasy service providers like ESPN, Yahoo, and CBS Sports introduced their own rankings and projections. More mock drafts.

By the time the average person who plays fantasy baseball got into a draft or auction room, they were looking at guidebooks and rankings based on collected intelligence gathering and making "choices."

Perhaps one individual may have decided to dismiss the hype on a certain sleeper. Perhaps another individual may have fixed himself to a player who was largely dismissed. But by and large, the fantasy baseball community arrived at a consensus valuation of every ballplayer in professional baseball.

As James Surowiecki pointed out in his book, "The Wisdom of Crowds," group decision-making, based on an aggregation of available data, tends to be surprisingly accurate compared to the decisions of a lone individual.

For this reason, it's very, very unwise to panic and make rash decisions on slumping or surging ballplayers based on what's happened to date in the 2009 season. Three weeks of baseball represents a small statistical sampling of less than 10 percent of the 2009 season. In most cases, the collected preseason wisdom of millions, who based their decisions on data collected and produced by a handful of experts, will still be the best oracle to what will happen in the final five months of the season.

However, there are some flaws in this group-sourced valuation process.

The first flaw concerns players without a great deal of experience in major league baseball. With a smaller body of work, young prospects are tough to project. Expert forecasters have tried to solve this by making use of "minor league equivalencies," but these numbers are still not as reliable as actual major league. In mock drafts, deviation on young, inexperienced players tends to be high. In other words, the consensus is less tight.

The second flaw concerns injuries. Nobody can predict them. Of course, some players are injury-prone (and the fantasy world adjusts valuations based on the tag), but any player dealing with a new, unexpected injury becomes less likely to hit a projection target.

The third and last flaw -- at least that we can think of at the moment -- concerns opportunity. Unfortunately, the fantasy baseball world holds no sway over the decisions of a real-life major league baseball manager. We may see a player as a good bet to reach 100 runs based on a high ability to get on base, but if a fickle manager disagrees and puts the batter last in his lineup, he'll also become more likely to disappoint. Conversely, a pitcher or batter who we've figured had little opportunity and then figures into a promotion, will surpass our expectations.

Consider these three things when deciding to veer off from the masses and make decisions based on a new "reality."