A major challenge for sportsbooks is "risk management" - setting lines to balance betting action. There are three schools of thought on sportsbook line management:
1. Market Clearing Model: the sportsbook tries to attract even money on each side of the wager. The losers essentially pay the winners and the sportsbook just collects the "vig".
2. Maximal Profits Model: the sportsbook takes advantage of bettor bias to increase its winnings. For instance, it may increase the spread on favorites knowing that the public is more likely to bet them. In effect, they are wagering on the unpopular side.
3. Accurate Pricing Model: the sportsbook tries to set a "correct" line. In the long term, they will profit by having the most accurate forecast of results, regardless of how much money comes in on either side of any one wager.
Traditionally, sports betting has been assumed to following a market clearing model where the sportsbook essentially just acts as a middleman, matching bets for a fee. More recent research on sports markets has challenged that idea, suggesting that sportsbooks may act to "take positions" to maximize profits. Understanding how the prices in a market are set can provide insight into "value" situations where advantages exist.
A recent seminar* entitled "Sportsbook Pricing and the Behavioral Biases of Bettors in the NHL" examined the betting market for games in the National Hockey League. This research, undertaken by Dr. Rodney J. Paul of St. Bonaventure University and Andrew P. Weinbach of Coastal Carolina University, used actual sportsbook betting data to uncover bettor bias in hockey markets.
Previous research on NFL and NBA betting markets has demonstrated a clear bettor bias for favorites in both sports. Teams with higher point spreads tend to receive a larger proportion of wagers; road favorites tend to be especially popular. However, a "contrarian" betting strategy (wagering against the betting public) has been profitable in the NFL, while it has not profitable in the NBA. This suggests that NFL markets might be following a maximal profits model, exploiting bettor bias, while the NBA is not. Instead, NBA point spreads seem to be based on an accurate pricing model, since favorites and underdogs performed similarly regardless of the amount of bets each side attracted.
Paul and Weinbach looked at sportsbook data for the NHL seasons from 2005-06 through 2007-08, the period since incorporation of the shootout rule. They found that road favorites were significantly overbet from what would be expected given the sportsbook odds. However, the betting odds were closer to the actual outcome of the games. In other words, bettors overpriced road favorites, but this was not taken into account by the sportsbook odds - the lines did not adjust to the bettor bias.
Examining some simple betting strategies revealed a few potential areas of profitability. It seems that, in general, road teams hold more betting value in the NHL. Betting on all road underdogs earned about 2 cents per dollar wagered, while betting only road dogs when their opponent was -200 or greater outperformed with 6 cents of profit per dollar wagered. Big road favorites (greater than -150) actually also showed a 3 cent profit across the board, despite receiving a much higher percentage of wagers than the betting odds would imply. Accordingly, betting on all home underdogs lost more than 3 cents per dollar wagered. The authors do validate a "contrarian" approach by noticing that sides that are more heavily overbet tend to perform more poorly, though sample sizes are small.
A possible explanation for the behavior of the NHL betting market is that public money influences the lines much more in the NFL than in the NBA or NHL. In other words, sharp bettors keep NBA lines "honest". The sheer volume of public bettors in the NFL allows the sportsbooks to relax their lines in order to exploit the public's betting bias. In the NBA or NHL, where public money is sparse, sportsbooks are forced to make accurate betting lines to prevent smart bettors from taking them to the cleaners. Paul and Weinbach suggest that this kind of pricing might serve to keep "wiseguys" out of the market.
Unfortunately the data available isn't sufficient to truly delve into "square" vs. "sharp" influence on the lines. The authors used percentage of wagers on each side, but more information can be provided by knowing the actual dollar amount on each side. Even better would be to know where the large bets went and where the small bets went to really look at the influence of "smart money".
Handicappers expecting to use this information should remember that markets are dynamic. The biases seen today may not be present next year as the market becomes more efficient. This is a particular concern for the NHL, as the market may be "immature" as it reacts to the implications of the new shootout rule and elimination of tie games. Handicapping immediately after the rule change involved a lot of unknowns- the market might be expected to self-correct after a few years of experience.
The bottom line: "fading the public" is generally a profitable strategy, but that may depend on the dynamics of the particular sports market. Even though the public prefers road favorites in the NHL, they are still profitable, and home underdogs actually perform poorly in the NHL despite being quite strong in other markets (i.e. the NFL). Knowing more precisely where the money is going on a game-by-game basis may uncover profitable opportunities. Most importantly, this underscores the importance of understanding the dynamics of the particular betting market in which you are active. Sportsbooks may use different pricing models depending on the market, and that may ultimately impact any profitable betting strategy.
*A preliminary draft of the authors' manuscript is available at http://www.uofaweb.ualberta.ca/economics2/pdfs/12-04-Seminar-Paul.pdf
If you would like to make or read comments about this article, you may do so by visiting the Mess Hall forum at MajorWager where a thread has been started. Please click HERE