William H. Dare

Efficiency in the NFL betting market: modifying and consolidating research methods

Modifying and consolidating previous research methods to generate more reliable estimates, some fairly weak evidence is found of inefficiency in the NFL betting market resulting from a bias favouring home underdog (against away favourite) teams. In contrast to previous research, no evidence is found that ‘momentum strategies’ generate significant returns in this market.

Re-examining the betting market on Major League Baseball games: is there a reverse favourite-longshot bias?

A recent study of the fixed-odds betting market on baseball games, while finding that the betting market is generally efficient, also found evidence of an underbetting on underdog teams. This article examines the evidence for this new anomaly. It corrects Woodland and Woodlands estimates of the commission, subjective win probabilities and test statistics. The efficiency null hypothesis cannot be rejected for all of their tests when revised test statistics are calculated for their sample period (however, like them, it was found that slight underdogs are underbet). It is also shown that their bias is not simply a bias involving favourites and underdogs. Whether underdogs are playing at home or away also seems to matter in their sample period. As well a positive relationship between returns and subjective probabilities was found for underdogs and favourites, a relationship suggestive of a favourite-longshot bias rather than its reverse. It is concluded that there is insufficient evidence to claim that this bias is a ‘true market inefficiency’.

The Search for Informed Traders in the Totals Betting Market for National Basketball Association Games

A recent article on line changes in the point spread betting market for National Basketball Association games found evidence that trading incorporates information into price. This article examines a closely related but previously unexamined betting market—the betting market for the total points scored in a game. The authors find that closing totals lines are more accurate forecasts of total points scored than are opening totals lines. It is shown that line changes move betting lines in the correct direction and by the appropriate magnitude to eliminate biases in opening totals lines. Line changes in this betting market, like those in the point spread betting market, cause prices to more accurately reflect fundamental values.

In search of the source of informed trader information in the college football betting market

The movement between the opening and closing of betting lines on sports events has been shown to contain valuable information. The purpose of this study is to search for the source of this valuable information. Changes in college football betting lines are examined with respect to information from the widely disseminated Dunkel Index, the Associated Press Writers’ Poll, and lagged performance variables. It is found that bettors use this information to bet but that this information is not significant to betting outcomes, which indicates noise trading. Most importantly, it is shown that the unexplained betting line movements are most important in explaining final game outcomes. This result leads to the conclusion that informed bettors are using private information.

A test for bias of inherent characteristics in betting markets.

The authors develop a model to investigate potential biases of inherent characteristics in betting markets. The test requires only that there be both a sides (“spread”) market and a totals (“over/under”) market for the game. The authors utilize the model to test for the well-documented “home-underdog” bias in the National Football League (NFL). They show that the bias specifically favors the offense (or defense) of home underdogs (away-favorites), with no bias against the offense (defense) of away favorites (home-underdogs).