Using data from the 4,858 baseball games that were played in the major leagues during the 2004 and 2005 seasons, four logit regression models that measure the likelihood of a team winning a game are estimated. Of particular interest is the effect of being the home team. As expected, the results indicate that a home-field advantage does exist in the major leagues, but only under certain circumstances. Specifically, the strength of the home-field advantage varies with the number of runs scored by the home team and with the run differential between the winning and losing team. The probability of a home team winning a game increases as it scores more runs, but it increases at a decreasing rate. Also, for a given number of runs scored, a home team is more likely to win a game than a visiting team. The home-field advantage is strongest in games where the run differential between the winning team and losing team is one run. It is weaker in games where the run differential is two runs and is non-existent in games where the run differential is three runs or more.