Young children's play is highly multimodal, with gesture, gaze, movement and speech often combined simultaneously in collaborative meaning-making. This article argues for a multimodal social semiotic perspective on play, recognising that this requires representation of data that brings multimodal elements into careful consideration. In this article, multimodal transcription is used to examine a video recording of three and four-year-old children playing a chasing game in an English nursery school. Map-like transcripts, including an animated transcript, are used to document an instance of their play, drawing particular attention to placement in space over time. Whilst such moments of play may at first appear fleeting and chaotic, multimodal transcription reveals the communicative, creative and agentive capacities of young children in a multitude of forms. The transcripts highlight and make evident the ways in which roles and rules of play are carefully negotiated moment-by-moment in multiple modes. In this way, map-like multimodal transcripts are presented as devices to highlight meaning-making where it may not normally be looked for, seen or recognised.