09/07/2010 09:38 pm ET Updated May 25, 2011

NFL Predictions: Ignore the Experts (except for me)

At long last, the NFL season is upon us. For most football fans, August is a time of unbridled optimism - every team is undefeated, everyone's fantasy football team is filled with sleepers, breakout players, and can't-miss prospects, and preseason games hint at the entertainment entailed during the regular season. Yet one particular type of football commentary tends to dominate all others in the days before the regular season: preseason predictions. Commentators in the sports media spend countless hours arguing about their preseason predictions for the upcoming season's outcomes. They do so with a zealot's conviction and certitude befitting a Nobel Laureate.

In my day job, I study the political economy of financial crises. Specifically, I study the effects of uncertainty in complex social systems, such as asset markets and the international system. All too often, I find that so-called experts in their fields cannot predict outcomes that seem obvious with hindsight. For example, after the crash of 2008, it became clear that the global financial system was a house of cards waiting to tumble down. Yet few economists identified these risks beforehand. Stock market "experts" like CNBC's bombastic Jim Cramer extol a psychic virtue of knowing when stocks are undervalued or overvalued - Cramer's stock picks then under-perform the market. In 1979, sociologist Ezra Vogel predicted that Japan's economy would surpass that of the United States' in size and productivity. This year, China surpassed Japan as the world's second largest economy.

No matter how hard the experts try, and no matter how certain they seem a priori, experts stink at prediction. Indeed, it seems like the only predictable thing about complex social systems is their unpredictability: there are just too many variables and feedback loops that make prediction an impossible task.

Back to football, just like the authorities in the social sciences stink at making predictions, so too do sports commentators of all stripes. Consider the following examples:

In 2008, sixteen of ESPN's self-described "experts" predicted that San Diego (5), Dallas (5), New England (3), Indianapolis (2), and Jacksonville (?!, 1) would win the Super Bowl. Pittsburgh won the Super Bowl that year. In 2009, experts predicted Pittsburgh (7), New England (5), and San Diego (4). New Orleans won.

Peter King, Sports Illustrated's senior NFL writer, predicted the Patriots would beat the Bears at the end of the 2009-2010 season. The Patriots were one-and-done in the playoffs; the Bears finished 7-9.

Prior to the 2009 season, ESPN's "Sports Guy," Bill Simmons, declared that the Buffalo Bills would finish with the worst record in football - they finished the season with six wins, better than eight other teams. Simmons asserted that Minnesota's Head Coach Brad Childress would be fired by the end of the season - he is still coaching today, and the Vikings made it to the NFC Championship game last year. Simmons maintained that the Arizona Cardinals would succumb to the Super Bowl "loser's curse" and wouldn't make the playoffs with a 5-11 record - they won a playoff game and finished the regular season with a 10-6 record.

What is true for the macro level, or the entire season, is also true for the micro level of individual games. The experts try to predict outcomes, but they are dismal. After starting on a hot streak, Simmons finished the year with a betting record against the spread of approximately 52% correct, slightly better than a coin flip. With a payout like that, you should have just bought Japanese stocks in 1980.

Of course, some experts are right sometimes. But I believe that these instances depend more on luck than supreme prescience. Otherwise, certain commentators could consistently predict a string of results over time. To my knowledge, no expert has accomplished that feat as of yet.

King and Simmons, who follow professional football day-in, day-out, offer no clairvoyance to the average sports fan. Yet their dearth of foresight is only surpassed by their popularity. NFL commentators, just like the players about whom they opine, are simply cogs in the NFL entertainment machine. Their attempts at prediction are another way in which the NFL tries to sell its product to its fan base. Validity is not their aim - entertainment is! From this perspective, King and Simmons are top-rate narrators. They sure can tell stories. But prediction? Forget about it.

Still, part of the fun of the NFL is trying to predict the outcomes of games. But the so-called experts are still just as uninformed as the average fan.

So, why is correct prediction such an elusive target? I see several reasons for this:

First, NFL games are random variables. If the 2007-2008 Super Bowl between the New York Giants and the New England Patriots were played one hundred times, I believe the Patriots would have won seventy-five of those match-ups. Perhaps a staunch believer in the Giants would weight the probabilities closer to fifty-fifty. Nonetheless, the game is only played once, so "on-average-I-would-have-been-right" analysis is less useful in football. In case you were wondering, the Giants won.

Secondly, commentators and fans alike are subject to availability bias, or overestimating the likelihood of an event associated with memorable occurrence. Because the Steelers won the Super Bowl in 2008, ESPN's experts overestimated the chance that the Steelers would win the Super Bowl in 2009. These experts likely understate the luck required to advance to the Super Bowl and overstate the continuity of inter-season momentum of winning teams.

Third, and by far the most important, is that uncertainty manifests itself in a plethora of ways in football. Many expected New England would bounce back from their 2007 Super Bowl loss - instead Brady was injured in the first game of 2008, and New England missed the playoffs. In 2005, the 14-2 Indianapolis Colts seemed poised to beat the visiting Steelers in the Divisional Round of the playoffs. Then, just prior to the game, Head Coach Tony Dungy's son committed suicide, distracting the Colts' leadership and surely undermining their home field advantage over the upstart Steelers. NFL history is filled with these unpredictable events: freak injuries to indispensable players, off-the-field troubles for ostensible model citizens, tuck rules, helmet catches, bobbled holds, daring onside kicks, favorable spots, and home run throwbacks (or, dare I say, forward passes?) abound.

But, not surprisingly, the things that make the NFL so unpredictable are also the things that make the NFL such a joy to watch. Who could have foreseen Tyree's helmet catch before the Super Bowl? Who could have imagined that a forgotten quarterback drafted in the sixth round from Michigan named Tom Brady would turn into a sure-fire Hall of Famer? This randomness and unpredictability is the fabric of the NFL. Without it, football would not be the same.

So my advice for you is simple: go ahead and buy into the hoax that we can predict outcomes in the NFL. Keep trying to predict the unpredictable. Pick a pair of Super Bowl contestants before the season starts and defend them for your dear life (how about the Texans and 49ers?) - this is the point of football. Just don't be too disappointed when you're wrong. You're in good company.