The purposes of this study are to examine the relation between playing commercial-off-the-shelf (COTS) games in the wild and L2 English vocabulary and to offer comparisons with non-gamers' vocabulary. Data were collected from two samples of teenage L2 English learners in Sweden, Sample A (N = 1,069) and Sample B (N = 16). Questionnaires and English grades were collected from A and B, productive and receptive vocabulary tests from A, and interviews and essays from B. A quantitative-dominant mixed-methods approach was adopted. Results showed a significant positive correlation between time played and test scores. They also showed that time played was related to types of games played. Multiple regression analysis including time played and types of games as predictor variables and L2 vocabulary as the outcome variable showed that the effect from type disappeared when it was entered into the model, whereas time remained significant. A close examination of 45 words (productive test) revealed significantly higher scores for gamers (compared with non-gamers) at all vocabulary frequency levels, and for particularly difficult words. Overall, findings from Sample B regarding gaming habits and vocabulary (i.e., use of advanced or infrequent words in essays) reflected the results from Sample A, making it possible to conclude that playing COTS games matters for L2 learner vocabulary.