A substantial body of research in the field of digital game-based language learning (DGBLL) suggests that games may hold significant potential to facilitate second language acquisition (SLA). Since many CALL researchers are also language teachers, it is common for studies in this field to be carried out in a language classroom context, with a focus on implications for in-class learning and teaching. Scholars such as Sauro and Zourou (2019) have recently pointed out, however, that DGBLL is far more likely to take place outside of the formal educational context than within it. Of the billions of digital game players around the world, many play in languages that are not their L1 and this must surely contribute significantly to their language learning. To better understand how SLA may take place in the context of a cooperative multiplayer digital game, a case study was conducted in which four hours of spoken interactions between three learners playing a game face-to-face over several sessions were recorded, transcribed, and analysed. Other than stipulating the use of English, the researcher did not guide or structure the interaction in any way, so as to simulate play ‘In the wild’ Interaction was analysed using two different qualitative approaches: a cognitive-interactionist analysis and a direct qualitative analysis. The first approach revealed occasional instances of negotiation for meaning and regular use of beneficial interactional strategies during interaction, while the second shed light on how unstructured game-based interaction among learners may facilitate the acquisition of L2 vocabulary and grammatical structures.