This study evaluates the predictive accuracy of expert (judgmental) forecasting and system (statistical) forecasting that provided two types of predictions on the NFL regular season games during 2000 and 2001 season. Then, we empirically investigate whether there exists a statistically significant difference in the predictive accuracy within the four groups of experts and systems respectively. From this, we implicitly investigate whether there is any prediction that has a better productivity. For this, we use the methodology of ‘Analysis of Variance by Ranks’ as an analytical tool. Our study finds that there is a statistically significant difference in the predictive accuracy within the group of experts as well as the group of systems who predicted the game winners. Meanwhile, we do not find a statistically significant difference within the group of experts as well as the group of systems who predicted against the betting line. In conclusion, it would not be the model of forecasting but the type of prediction that may induce a difference in the predictive accuracy.