Objective: The self-isolation measures employed during the COVID-19 pandemic made it difficult for basic needs to be met, thus increasing emotional distress. It has been suggested that socially meaningful online gaming buffered emotional distress during the lockdown. This study aimed to test the protective effect of online gaming during the lockdown and to investigate the differences between highly involved gamers (those who play videogames intensely without adverse consequences) and problematic gamers in this regard. Method: Capitalizing on a data collection that started before the pandemic, we adopted a cross-sectional between-groups study design in which gaming patterns, gaming related variables and levels of emotional distress were compared between a pre-COVID group (N=298) and a COVID group (N=366). Results: Compared to the pre-COVID group, high involvement in gaming was more prominent and emotional distress was decreased in the COVID group. Moderated regression analyses further revealed that the interaction between social compensation via gaming and highly involved gaming was associated with lower levels of emotional distress in the COVID-19 group. In contrast, the interaction between gaming-related relaxation and problematic gaming predicted higher emotional distress in the COVID-19 group. Conclusions: This study suggests that gaming for social compensation might mitigate the experienced emotional distress during pandemic related self-isolation, whereas maladaptive gaming patterns could constitute a vulnerability factor deserving clinical attention.