Optimizing the Performance of Expansion Teams with Application to the NHL's Vegas Golden Knights

Abstract

In the past across all sports expansion teams have come into the league hoping for early success especially in their first year. This success has eluded most of these expansion teams, especially the past four NHL expansion teams. The Vegas Golden Knights are the next NHL expansion team and they are entering the league at the start of the 2017-18 season. This study used statistical analysis to determine the optimal selection of players for the Vegas Golden Knights. A set of eight possible statistically optimal teams was collected from two different available player lists. From there the 16 teams were put through a neural network to predict the outcome they would have over a 10 game set. The neural network was trained and tested from 930 10 game sets over the past four years. When the predicted optimal team was found it was then used in a neural network to predict the outcomes of singular games. This neural network was trained and tested with 990 games form the 2016-17 season. The findings of this study are an important step towards the use of statistical analysis in hockey and can provide useful tools for general managers.