Executive Report Analytical Tools

EXECUTIVE REPORT

National Hockey League

TABLE OF CONTENTS

Summary1

Introduction1

Possible Factors Affecting the Prediction1

Building a Multiple Regression Model1-2

Testing the Multiple Regression Model2-3

What are the best predictors?2

Why were the models chosen? 3

What is the best model?3

Conclusion3

Works Cited4

NHL Team Winning Prediction Executive Report

SUMMARY

The National Hockey League (NHL) is an association which operates 30 professional ice hockey teams. The NHL would like to know, based on last year’s season information, what the team winning predictions will be for the upcoming season. Using that data, we created and tested data using several methods to determine the most effective combination.

INTRODUCTION

The features within these regression models rely on data from the 2011-2012 season. The process of using the regression models was to determine the relationship among variables chosen. The variables that were used were, penalty kill percentage, goals per game, goals against per game, 5 on 5 for/against ratio, shots against per game, shots per game, and power play percentage. Here, regression models were used to predict team wins during the 2012-2013 season. The model that was chosen had the most significant relationship of the independent variable(s) beyond the two dependent variables: (a) goals scored per game (b) goals scored by opposing team per game.

POSSIBLE FACTORS AFFECTING THE PREDICTION 

a) Penalty Kill Percentage

b) Goals Scored Per Game

c) Goals Scored by Opposing Team per Game

d) 5-on-5 Scoring for/against Ratio

e) Shots Taken by Opposing Team per Game

f) Shots Taken During Game

g) Power Play Percentage

BUILDING A MULTIPLE REGRESSION MODEL

Upon building the multiple regression models, the importance of dependent and independent variables were chosen based on recorded statistics.

In this project, the following variables are dependent variables:

* Goals Scored...