The first set of regression results uses the number of bets placed on each game as the dependent variable in the regression model described by equation (11.1). Again, this variable is the total number of bets placed at the four online sports bookmakers tracked by Sports Insights. Independent variables include an intercept, the months of the year, the days of the week, dummy variables for elite pitchers (both visitor and home), the absolute value of the odds on the game, a dummy for road favorites, and the win percentages of both the home and road teams going into the game.
The months of the baseball season are included to control for any seasonality in betting volume. This seasonality could manifest itself as bettor excitement early in the year, effects of the start of football betting on baseball betting volume at the end of the season, or other reasons. The omitted monthly category is the first month of the season, April. Weekdays are also included as dummy variables to account for possible daily variation in betting volume.
The dependent variable is betting volume per game, so there could be impact where bettors may place more wagers on days when there are fewer games (especially when we consider baseball gambling as purely an activity of consumption). Mondays and Thursdays typically have fewer baseball games, as these serve as travel days 200 sports betting for teams. If bettors choose to wager on baseball as a form of consumption, due to fewer games on these days, individual game betting volume may increase on Mondays and Thursdays.
Weekend effects also are possible, as more bettors may engage in gambling on the weekend as compared to weekdays. Tuesday is the omitted category of the days of the week. The“Cy Young Voted Pitcher” variable (both home and road) is the proxy for the elite pitchers as mentioned in the introduction. The top five starting pitchers who received votes for the Cy Young Award in each league (NL and AL) are given a value of one, while all other pitchers are given a value of zero.
The first regression specification uses pitchers from the previous season (2008) who met this criterion, while the second uses the current year to create the dummy variable, and the last specification combines the two years (if a pitcher appeared once or twice on the list he is given a value of one).
We will use the previous season dummy as the main point of discussion of the results, but if one believes that elite pitchers can easily be identified during the season, new elite pitchers are likely to receive attention as the season progresses. Although this is a simple version of identifying an elite pitcher, we believe it will give a good representation of the importance of the pitcher in the minds of bettors of Major League Baseball.