One Way to Make Money Betting on Football

Back in 2004, University of Chicago econ professor and Freakonomics blogger Steven Levitt showed betting on home underdogs in the N.F.L. is a consistent winning strategy.

What is a home underdog? A typical betting line might be spoken as: "New England minus four at Houston." This means Houston is the home team and is expected to lose to New England by four points, hence a home underdog. If New England doesn't beat Houston by more than four points then bets placed on Houston win. (The case where New England wins by exactly four points is called a "push" and all bets are returned to the bettors.)

Levitt used results from the 2001-02 N.F.L. season to show that bets on home underdogs had a success rate of 58 percent.

Now, a new paper takes a fresh look at the phenomenon and finds that the home underdog effect takes place almost exclusively in the latter parts of a season. (free version)

Richard Borghesi, an economist at Texas State University, looked at N.F.L. games between 1981 and 2000 and concludes that if late season games and playoff games are completely removed (accounting for 18 percent of all games), the bias disappears.

The graph below shows success rates, excluding pushes, on betting on the home underdog during the entire regular season versus just late in the season and the playoffs:

And this graph shows the breakdown between two simple betting strategies: Betting on the home team and betting on the home underdog, at different parts of the season:

One source of the bias is that bettors don't fully take into account the role of weather, according to Borghesi. With most teams playing in open-air stadiums, it's been argued that away teams from mild climates are at a disadvantage when they play in cold regions in the latter parts of the regular season and playoffs.

Borghesi looked at 200 games played in cold climates after week 14 and found the average closing line predicted home teams would win by 3 points, but in reality those teams won by 7 points.

Another source for the bias could be that as the season progresses, increased media attention helps bring in a lot of "dumb money." Betting statistics in Las Vegas show that volume does increase sharply in the latter parts of the N.F.L. season.

For his part, Levitt attributed the bias to the ingenuity of market makers (i.e. bookies): Bookies aren't merely happy with moving the betting line so that the dollar amount on one side is equal to the dollar amount on the other -- in this situation the bookies just collect the 'vig', or commission.

These market makers, Levitt said, are probably better at predicting the final outcome of games, so it's in their interest to position the spread such that there are more losers than winners. In that case, the extra money after all the winners have been paid off has nowhere to go but in a bookie's pocket.

Now, I lied a little with the title to this post. Levitt and his co-author Dubner wrote about this phenomenon for a certain widely read publication in February 2006, so you might assume the arbitrage opportunity of betting on home underdogs is dead. But still, Levitt tried to use it for the 2006 - 2007 season seems to have achieved relatively successful results.

Related:

• Why Are Gambling Markets Organised So Differently From Financial Markets? (pdf)
• The Home Team Weather Advantage and Biases in the NFL Betting Market ($)
• Cafe Hayek's Critique of Home Underdog Advantage