Stock investors are not fully rational during their trading, and many behavioral biases affect their trading behavior, such as representative bias and disposition effect. However, most of the literature on behavioral finance cast efforts on explaining empirical phenomena observed in financial markets, but little on how individual investors' trading performance is affected by their behavioral biases. As against the common perception that behavioral biases are always detrimental to investment performance, we conjecture that these biases can sometimes yield better trading outcomes for investors. Focusing on representative bias and disposition effect, we construct a mathematical model in which the representative investor follows a Bayesian trading strategy based on an underlying Markov chain, switching between Trending regime and Mean-reversion regime. By this model, we are able to undertake scenario analysis to track investor behavior and performance along the time, under different patterns of market movements. Results validate our conjecture by showing that the effect of behavioral biases can sometimes be positive on investor performance.