One of the key trends in learning technology in recent years has been the growing interest in simulations and serious games. Much of the research into simulations and serious games has tended to focus on aspects of fun and engagement rather than on what is happening in the process to cause transferable learning. Much of it has been based on small samples with results that show significant levels of acceptance and enthusiasm from the learners, but the real learning gains are harder to discern. As is so often the case in learning technology, potentially valuable underpinning research lies forgotten and unused. The aim of this article is to dust off some earlier research reports to see what metrics we can use, and experiments we can carry out, to help us understand how it is that people learn from simulations and perhaps serious games--and so help us to design materials that are more effective.