Rationale: Assessing complex problem-solving skills (CPS) is of great interest to many researchers. However, existing assessments require long testing times making them difficult to include in many studies and experiments. Here, we propose a specific composition of microworlds based on the MicroDYN approach, which allows for valid estimation of CPS in a substantially reduced amount of time (<20 min). Methods: Based on the reanalysis of a sample of N = 232 university students who worked on 11 microworlds of increasing difficulty, we conducted multiple confirmatory factor analyses to test all possible combinations of microworlds, which were theoretically justified in advance. Results/Discussion: We demonstrate one best fitting set with five microworlds, which shows excellent factorial validity and relates to both conventional measures of intelligence and to school grades. We hope that this will allow other researchers to include CPS into their study designs even when testing time is limited.