Experimental work has demonstrated that the efficient operation of the GroEL-GroES chaperonin machinery is sensitive to the presence of macromolecular crowding agents. Here, I describe atomically detailed computer simulations that provide a microscopic view of how crowding effects are exerted. Simulations were performed to compute the free energy required to extract the protein rhodanese from the central cavity of GroEL into solutions containing a range of crowder concentrations. The computed energetics allow the total yield of folded protein to be predicted; the calculated yields show a nonlinear dependence on the concentration of crowding agent identical to that observed experimentally. The close correspondence between simulation and experiment prompts the use of the former in a truly predictive setting: simulations are used to suggest that more effective crowding agents might be designed by exploiting an "agoraphobic effect."