For example. Kesten C. Green has a great paper on the efficacy of role playing vs. game theory vs. individual assessment for conflicts involving small numbers of parties with a lot at stake. He finds that role playing is the only forecasting method he tested that actually predicts what will happen.
Green did a study of how well forecasting methods successfully predict the outcome of conflicts involving small numbers of people with lots at stake. He discovered that only role playing actually predicted better than guessing.
So why aren’t we role playing more often? Cult of the Expert?