The purpose of this paper is to evaluate the accuracy of traditional statistics and advanced statistics at predicting player salaries in the National Hockey League (NHL). I wanted to find out if advanced statistics could add more accuracy to the basic statistics model. I used OLS multivariate regression analysis on three separate models; one with only basic statistics, one with only advanced statistics, and one with both basic and advanced statistics. Additionally, the sample of players was split up into groups of forwards, defensemen, unrestricted free agents, and restricted free agents. Through comparing the basic statistic and advanced statistic models, basic statistics were found to be more accurate in predicting salaries. Some of the statistics that had the most statistical significance include average time on ice and penalty minutes. While advanced statistics were not terrible predictors, they did not contribute much extra goodness of fit when they were included in the combined model. By comparing the different groups that the sample of players were split up into, I found that the model did a better job of predicting defensemen over forwards and restricted free agents over unrestricted free agents. From my results, I found that it is plausible that most contracts in the NHL have been signed without much consideration of advanced statistics as general managers do no pay attention to them when negotiating with players. This could be due to the fact that advanced statistics are still relatively new to the league.