I was surfing the Internet when I stumbled upon Brian Macdonald's paper from the MIT Sloan Sports Analytics Conference, which details a model to calculate expected goals for players and teams:
I figured I could translate this expected goals model into game ratings by taking each variable and calculating the regression formula on a single-game basis.
For example, let's look at Winter Classic ratings:
Boston
For example, let's look at Winter Classic ratings:
Boston
Name | Goals | Shots | Hits | Hits Against | Faceoffs | Rating |
Jimmy Hayes | 0 | 1 | 3 | 1 | 1 | 0.01 |
Brett Connolly | 0 | 2 | 1 | 0 | 0 | 0.03 |
Loui Eriksson | 0 | 1 | 1 | 0 | 0 | 0.00 |
Max Talbot | 0 | 0 | 0 | 1 | 18 | 0.75 |
Landon Ferraro | 0 | 1 | 3 | 0 | 8 | 0.26 |
Zdeno Chara | 0 | 1 | 1 | 0 | 0 | 0.00 |
Zac Rinaldo | 0 | 0 | 3 | 1 | 0 | -0.06 |
Patrice Bergeron | 0 | 4 | 1 | 1 | 17 | 0.80 |
Matt Beleskey | 1 | 4 | 5 | 1 | 0 | 0.31 |
Dennis Seidenberg | 0 | 0 | 2 | 3 | 0 | 0.03 |
Joe Morrow | 0 | 3 | 1 | 2 | 0 | 0.12 |
Torey Krug | 0 | 1 | 0 | 2 | 0 | 0.09 |
Ryan Spooner | 0 | 1 | 0 | 1 | 17 | 0.74 |
Seth Griffith | 0 | 0 | 0 | 1 | 0 | -0.03 |
Adam McQuaid | 0 | 0 | 3 | 1 | 0 | -0.06 |
Frank Vatrano | 0 | 0 | 0 | 1 | 0 | 0.03 |
Alexander Khokhlachev | 0 | 0 | 1 | 0 | 1 | 0.01 |
Kevan Miller | 0 | 1 | 2 | 1 | 0 | 0.00 |
​Montreal
Name | Goals | Shots | Hits | Hits Against | Faceoffs | Rating |
Brendan Gallagher | 1 | 3 | 0 | 2 | 0 | 0.46 |
Tomas Plekanec | 0 | 1 | 2 | 2 | 12 | 0.51 |
Tomas Fleischmann | 0 | 1 | 0 | 0 | 0 | 0.03 |
Torrey Mitchell | 0 | 0 | 1 | 0 | 10 | 0.37 |
Dale Weise | 0 | 3 | 0 | 0 | 0 | 0.18 |
Jeff Petry | 0 | 3 | 0 | 2 | 0 | 0.15 |
Alex Galchenyuk | 0 | 1 | 1 | 0 | 14 | 0.56 |
Nathan Beaulieu | 0 | 1 | 1 | 2 | 0 | 0.06 |
Brian Flynn | 0 | 1 | 0 | 1 | 5 | 0.26 |
Paul Byron | 2 | 2 | 0 | 1 | 0 | 0.71 |
Daniel Carr | 0 | 1 | 5 | 2 | 0 | -0.06 |
Mark Barberio | 0 | 2 | 0 | 3 | 0 | 0.15 |
David Desharnais | 1 | 2 | 2 | 1 | 16 | 0.98 |
Max Pacioretty | 1 | 1 | 0 | 2 | 0 | 0.40 |
Alexei Emelin | 0 | 3 | 1 | 0 | 0 | 0.06 |
P.K. Subban | 0 | 1 | 3 | 4 | 0 | 0.06 |
Andrei Markov | 0 | 0 | 2 | 3 | 0 | 0.03 |
Lars Eller | 0 | 0 | 3 | 1 | 3 | 0.06 |