Document Type


Publication Date


Department 1



Each July, the eyes of baseball fans across the country turn to Major League Baseball’s All-Star Game, gathering the best and most popular players from baseball’s two leagues to play against each other in a single game. In most sports, the All-Star Game is an exhibition played purely for entertainment. Since 2003, the baseball All-Star Game has actually ‘counted’, because the winning league gets home field advantage in the World Series. Just one year before this rule went into effect, there was no winner in the All-Star Game, as both teams ran out of pitchers in the 11th inning and the game had to be stopped at that point. Under the new rules, the All-Star Game must be played until there is a winner, no matter how long it takes, so the managers need to consider the possibility of a long extra inning game. This should lead the managers to ask themselves what the probability is that the game will last 12 innings. What about 20 innings? Longer?

In this paper, we address these questions and several other questions related to the game of baseball. Our methods use a variation on the well-studied geometric distribution called the quasigeometric distribution. We begin by reviewing some of the literature on applications of mathematics to baseball. In the second section we will define the quasigeometric distribution and examine several of its properties. The final two sections examine the applications of this distribution to models of scoring patterns in baseball games and, more specifically, the length of extra inning games.

Required Publisher's Statement

Original version is available from the publisher at: