Understanding the driving forces in the markets of their products is a basic necessity for any business. Quantitative models are either aggregated over large market segments or restricted to utility models of an individual's buying decision. While the aggregate models acknowledge that customer interactions are important they do not model them and hence have no way to adjust their model to changing business environments. This paper bridges the gap between individual decisions and the overall market behavior using agent based simulations to model the sales of computer chips in the high-end gamers market. The simulation environment is dynamic and models the succession of 19 products introduced over a 40 month time horizon which includes the recession of 2008-2010. Simulated sales are compared to actual sales data and are used to adjust the parametrization of the agents and their environment. We found that only two agent parameters are sufficient to obtain a very reasonable fit between simulations and data: The amount of money available for the gaming hobby and a parameter related to the gaming success of the high-end gamers.