Biased Decision Making In A Naturalistic Environment: Implications For Forecasts of Competitive Events