The model simulates every matchup 10,000 times to determine the statistical probability of specific outcomes, such as hitting the 'Over/Under' or covering the point spread. By running these iterations, it identifies discrepancies between its own aggregate projections (e.g., projected points) and the lines established by commercial betting markets.