Aims: Risky single-occasion drinking (RSOD) by young people is a serious public health issue, yet little is known about the specific circumstances of risky drinking occasions. This study examined the independent effects of event- and individual-specific variables on RSOD. Design: Longitudinal cohort study measuring self-reported RSOD and event- and individual-specific variables across two drinking occasions approximately 1 year apart. Setting: Metropolitan Melbourne, Australia. Participants: A sample of 710 young risky drinkers aged between 18 and 25 years and defined as engaging in risky drinking practices (males: consumed alcohol in excess of 10 Australian Standard Drinks (ASD: 10 g ethanol) in a single occasion in the previous year; females: consumed alcohol in excess of seven ASD for females in a single occasion in the previous year). Measurements: Random digit-dial telephone landline survey of the most recent heavy drinking occasion and socio-demographic variables. The primary outcome was the log of the total drinks consumed in the most recent heavy drinking occasion. Event-specific (e.g. number of drinking locations) and time-varying (e.g. weekly income) and time-invariant (e.g. sex) individual-specific variables were examined as correlates of total drinks consumed. Findings: Changes in event-specific characteristics including the length of the drinking occasion (Likelihood Ratio χ2(2) = 24.4, P < 0.001), the number of drinking locations (Wald χ2 (1) = 7.6, P = 0.006) and the number of different drink types (Wald χ2 (1) = 13.6, P < 0.001) were associated with increases in total drinks consumed, after adjustment for time-invariant and time-variant individual-specific variables such as gender, income level and weekly consumption. Few other effects were noted. Conclusions: Event-specific characteristics are important predictors of the number of drinks consumed during risky single occasion drinking (RSOD) and illustrate the importance of event contexts when considering interventions targeting RSOD. The total number of drinks consumed in a RSOD session appears to rise independently with the duration of the drinking event, the number of drinking locations and the number of different types of beverage consumed.