As the worldwide popularity of Internet gambling increases, concerns about the potential for gambling-related harm also increase. This paper reports the results of a study examining actual Internet gambling behavior during 10 years of play. We examined the electronic gambling records of subscribers (N = 2,066) who triggered a responsible gaming alert system at a large international online gaming company. We compared these cases with control subscribers (N = 2,066) who had the same amount of exposure to the Internet gambling service provider. We used discriminant function analysis to explore what aspects of gambling behavior distinguish cases from controls. Indices of the intensity of gambling activity (e.g., total number of bets made, number of bets per betting day) best distinguished cases from controls, particularly in the case of live-action sports betting. Control group players evidenced behavior similar to the population of players using this service. These results add to our understanding of behavioral markers for disordered Internet gambling and will aid in the development of behavior-based algorithms capable of predicting the presence and/or the onset of disordered Internet gambling.