The accuracy and influence of financial analyst forecasts have drawn significant attention due to their essential role in guiding investment decisions and shaping market sentiment. Earnings per share (EPS) forecasts, in particular, provide critical insights into corporate performance, influencing equity valuations and market trends. While EPS is a widely used measure of profitability and financial health, research highlights persistent biases that undermine its reliability. These biases often arise from information asymmetry, behavioral tendencies such as herding and overconfidence, and potential conflicts of interest, leading to systematic forecast errors. The I/B/E/S database aggregates detailed analyst data, including earning forecasts for publicly traded companies. This study evaluates analyst performance through a dynamic ranking system that measures EPS forecast accuracy over time. By periodically ranking analysts, we identify high- and low-performing forecasters while assessing the stability of their predictions. To improve forecast accuracy, we introduce an enhanced consensus method that surpasses individual estimates by applying rank-based weighting. Our approach leverages iterative filtering algorithms to refine consensus estimates by computing key parameters such as individual and market variances and a reliability index. These metrics are integrated into a composite score, allowing for adjustments to prediction discrepancies and the identification of long-term reliability patterns, ultimately improving the accuracy and robustness of EPS consensus forecasts.