Goal Prevention 2004 a Review of Goaltending and Team Defense Including a Study of the Quality of ’
Goal Prevention 2004 a review of goaltending and team defense including a study of the quality of a hockey ’shots allowed Copyright Alan Ryder 2004 Goal Prevention 2004 Page 2 Introduction I recently completed an assessment of “”in the NHL for the 2002-03 “”season (http://www.HockeyAnalytics.com/Research.htm). That study revealed that the quality of shots allowed varied significantly from team to team and was not well correlated with the number of shots allowed on goal. The consequence of that study was an improved ability to assess the goal prevention performance of teams and their goaltenders. This paper applies the same methods to the analysis of the 2003-04 “”season, focusing more on the results than the method. Shot Quality In summary, the approach used to assess the quality of shots allowed by a team is: 1. Collect, from NHL game event logs, the relevant data on each shot. 2. Analyze the goal probabilities for each shooting circumstance. In my analysis I separated certain “”from “”shots and studied the probability of a goal given the shot type, the ’distance and the on-ice situation (power play vs other). 3. Build a model of goal probabilities that relies on the measured circumstance. 4. Apply the model to the shot data for the defensive team in question for the season. For each shot, determine its goal probability. 5. Determine Expected Goals: EG = the sum of the goal probabilities for each shot. 6. Neutralize the variation in the number of shots on goal by calculating Normalized Expected Goals (NEG) = EG x League Average Shots / Shots 7.