A criticism of current diagnostic approaches to gaming disorder is that they fail to take into account that high and repeated engagement is not problematic per se, nor is it necessarily associated with adverse consequences. To tackle this controversy, we used Confirmatory Factor Analysis (CFA) to test, in regular gamers (N = 268), whether high (but healthy) engagement can be distinguished from problematic engagement by using the Addiction-Engagement Questionnaire (Charlton & Danforth, 2007). We then tested whether differential relationships exist between the engagement and addiction constructs, DSM-5 criteria for Internet gaming disorder (IGD), and psychological factors linked to gaming use and misuse (self-reported impulsivity, motives to play, and depression). Results indicated that a model holding engagement and addiction as two distinct, but related, constructs fits the data well. Second, we showed that although both constructs are linked to the number of IGD criteria endorsed, the relationship is more pronounced for the addiction construct. Third, a differential pattern of correlations was observed with the other study variables, further supporting the need to distinguish the two constructs. Our study emphasizes that research is needed to refine the diagnostic approach to gaming disorder to avoid conflating healthy passion with pathological behavior.