The presence of others is assumed to intensify gambling behaviour. However, empirical findings are mixed. Against this backdrop, a gambling simulation experiment was designed, comprising three conditions (gambling alone, gambling with two unfamiliar others and gambling with two familiar others). A total of 136 participants were recruited and assigned to one of the conditions. Win-sequence (‘early’, ‘middle’ and ‘late’) was a sub-condition. The gambling simulation was based on the electronic gaming machine (EGM). The dependent repeated measures were ‘Total trials gambled’, ‘Bet size’, ‘Reaction time’ and ‘Credit’. Several univariate mixed regression models were fitted. Contrary to expectations, participants gambling alone gambled on average more trials, placed their bets faster and lost fewer credits over time compared to those gambling with other participants. Participants receiving reinforcements mainly during the first part of the session showed more persistence within the gambling session in comparison with participants receiving winnings mainly at a later stage of the session. This suggests that the arousal induced by fellow participants facilitated a standard of maximizing actual returns, resulting in a loss-avoidance strategy. As attitudes toward gambling in Norway are relatively negative, the results could also reflect normative group pressure or fear of negative evaluation from others.