Behind Hockey's Stars: a Statistical Analysis of Critical Hockey Team MetricsMain MenuAuthor's NotesBackground InformationMethodologyA description of the tools used to create my modelThe Final ModelConclusionsThe key takeaways when using the modelFurther ImplicationsBibliographyRennie Hezae1990fe70b4748351393d4d28793c01835984608
Keith Slash
12017-06-28T20:55:27-07:00Rennie Hezae1990fe70b4748351393d4d28793c01835984608197941Duncan Keith earns a penalty in the 2014 playoffsplain2017-06-28T20:55:27-07:00Rennie Hezae1990fe70b4748351393d4d28793c01835984608
The fourth metric in the model is the percentage of "penalties killed" by the team. A team becomes "short handed" when a player takes a penalty. Taking a penalty refers to breaking any of hockey's major rules: no tripping, no late hits, no hitting the opposing goalie, etc. An example of a slashing penalty is shown to the right. When a penalty is taken, the team must player for two minutes with only four players as comparedto the opponents five players. In order to kill a penalty, the team must not allow a goal in those two minutes. Including this metric is intended to account for a team's ability to play smartly and efficiently, because playing with one less skater is draining and the offensive pressure of the opposing team is unrelenting. Interestingly, the number of penalties scored in a game seemed to have no relationship with the number of regular season points earned, but the penalty killing percentage did. This trend is clear below, though less strong than the trends shown above. The R-squared value is 21%. This is high enough to deem this relationship statistically significant because 21% of the variation in regular season point totals is related to penalty kill percentages.
Mouse over each datapoint to see more about specific teams, and toggle between playoff teams and non playoff teams to see what sets them apart! <p>&nbsp; </p>
