Tested the Elshout-Raahein hypothesis that the correlation between intelligence and problem solving in computer simulations is moderated by domain knowledge. Experiment 1 gave 100 college students and 100 secondary school students (aged 12-15 years) the "Hunger in the Sahel" computer simulation game under 4 different levels of preknowledge. Problem-solving achievement was measured in terms of effectiveness (number of expert decisions) and efficiency (length of survival in game). Intelligence was assessed with a subtest of the Berliner Intelligenzstruktur-Test. Experiment 2 gave the simulation to 28 young adults (mean age 20 years) in a within-subjects design varying preknowledge through growing game experience over 4 days. Performance was measured with efficiency. Results broadly confirmed the hypothesis. It is concluded that no new type of intelligence needs to be postulated to manage complex dynamical systems. Recommendations are deduced for the use of computer simulations in discovery learning and when assessing intelligence and problem-solving ability.