[Blume and Easley (1992)] show that if agents have the same savings rule, an expected discounted logarithmic utility maximizer with correct beliefs will dominate. If no agent adopts this rule, then agents with incorrect beliefs, but equally averse to risk as logarithmic utility maximizers may eventually hold more wealth than the agent with correct beliefs. In other words, a trader with correct beliefs can be driven out of the market by traders with incorrect beliefs. However, [Sandroni (2000)] shows that, among agents who have the same intertemporal discount factor and who choose savings endogenously, the most prosperous will be those making accurate predictions. Agents with inaccurate predictions will be driven out of the market regardless of their preferences. By using the extended agent-based articial stock market, we simulate the evolution of portfolio behavior, and investigate the characteristics of the long-run surviving population of investors. Our agent-based simulation results are largely consistent with [Blume and Easley (1992)], and we conclude that preference is the key factor determining agents' survivability.