Electronic gambling offers the opportunity to analyze huge and unbiased data sets of automatically recorded actual gambling behavior. This study refers to data on 2,127,887 poker playing identities from the Online Poker Database of the University of Hamburg (OPD-UHH) to analyze three subgroups of gamblers: regulars, newcomers, and dropouts. Their gambling habits over 6 months are analyzed in total, as well as over time. Regulars show a much higher involvement than non-regulars and increase their playing volume slightly over the observation period. Newcomers have a lower involvement than non-newcomers and most of them decrease their playing volume over time. Still, there is a small group of newcomers which increases their playing volume sharply and is, hence, very interesting for the industry as well as for the early prevention of pathological gambling. Dropouts have a higher gambling involvement than newcomers but play less than players who have not stopped stop gambling. Most dropouts also show a decreasing playing volume before dropping out. An analysis of the correlations between different variables of gambling habits shows that most of them reinforce each other, for example: gamblers with a higher total playing time tend to play at more tables simultaneously. Only playing frequency is a moderating variable of gambling involvement.