While most pathological gamblers, like most alcohol abusers, recover on their own, it seems likely that they are at greater risk for relapse than those who have been through successful treatment. Accordingly, a substantial increase in treatment resources for pathological gamblers, along with greater efforts to establish the effectiveness of these treatments, ought to receive national priority. If the data on alcohol-abusing self-changers are generalizable to self-changing pathological gamblers, the prognosis for gamblers who stop gambling all together is better than for those who aspire to controlled or nonproblem gambling. While pathological gamblers with comorbid substance abuse are more difficult to treat than those without it, the impact of comorbid substance abuse on the decision by pathological gamblers to change has not yet been explored, although it should be. Similarly, the impact of other Axis I pathology on pathological gambling self-change, especially depression and anxiety, should be thoroughly explored. The most pressing problem in this field appears to be definitional. As a consequence, until consensus is reached on a reliable, valid, and useful classification scheme for pathological gambling, both research and clinical efforts will continue to suffer.