Background: It has repeatedly been reported that when making decisions under uncertainty, groups outperform individuals. In a lab setting, real groups are often replaced by simulated groups: Instead of performing an actual group discussion, individual responses are aggregated by a numerical computation. While studies typically use unweighted majority voting (MV) for this aggregation, the theoretically optimal method is confidence weighted majority voting (CWMV) -- if confidence ratings for the individual responses are available. However, it is not entirely clear how well the theoretically derived CWMV method predicts real group decisions and confidences. Therefore, we compared simulated group responses using CWMV and MV to real group responses. Results: Simulated group decisions based on CWMV matched the accuracy of real group decisions very well, while simulated group decisions based on MV showed lower accuracy. Also, CWMV well predicted the confidence that groups put into their group decisions. Yet, individuals and real groups showed a bias towards under-confidence while CWMV does not. Conclusion: Our results highlight the importance of taking into account individual confidences when investigating group decisions: We found that real groups can aggregate individual confidences such that they match the optimal aggregation given by CWMV. This implies that research using simulated group decisions should use CWMV and not MV.