Using the PISA surveys (2000-2012), this paper explores the relationship between math test scores and everyday computer gaming by gender and for high income and middle income countries. We use two identification strategies in the spirit of an ideal experiment that would reduce computer gaming through limited internet access or through schools alternative demands. We find that everyday computer gaming has positive effects for boys, but negative effects for girls arising mostly in collaborative games suggesting a role for social effects. Computer gaming is becoming the new "swimming upstream" factor in the quest to close the gender gap in math.