ECHL Game Score

Back in 2016, Dom Luszczyszyn released Game Score. It was a metric originally introduced by the baseball phenomenon Bill James which showcases how a player did in each game. 

It was a completely different way to calculate it, as Bill James just used simple counting, and Dom used statistics with weights attached to each one. That has become the norm in hockey since, as anyone who has replicated it to their own means has done the same thing as Dom. 

Two years later, an article by the same website posted by Shawn Ferris released Game Score for the NWHL, the women’s hockey league. It was more toned down as the NWHL didn’t have access to statistics like the NHL did at the time. 

This formula was more simple than the one Dom put out. Dom eventually advanced his formula more, as more advanced analytics came to light. 

Recently, I stumbled upon an article by Bailey Johnson who showcased her Game Score formula for NCAA D1 Hockey. Which was a little more complex than the one Ferris put out, but still relatively simple for the normal everyday person. 

This gave me the idea to try something like it for my local pro hockey team, the Cincinnati Cyclones of the ECHL. 

Looking at the stats that the ECHL puts out, it isn’t much. The ECHL focuses more on the type of goal (PP, SH, Unassisted, Empty Net, Shootout) than other stats. The main stats that the ECHL puts out are: Goals, Assists, Points, Plus-Minus, Shots on Goal, and Penalty Minutes.

The ECHL doesn’t calculate much more than this, unless it just isn’t public to the fans. This probably has to do with the lack of money the ECHL generates, compared to other leagues like the AHL and NHL. 

The ECHL is known as the “AA of hockey”, as it is the 3rd tier of professional hockey, like AA baseball is to the MLB. 

So, in order to work with the lack of stats that we have, I used a combination of both Ferris’ and Johnson’s formulas in order to make this work.

Player formula: G+(0.64*A)+(0.11*SOG)+(0.09*+/-)-(0.07*PIM)

I messed around with some of the variables and this was one of the best I got. It did well with an R2 on both the Game Score compared to Points, and the Game Score per Game compared to Points per Game

Goalie formula: (0.16*SV)-GA

The goalie formula is about the same for most of the Game Score calculations out there. I messed around with the coefficient for the saves a bit in order to get a good number so that it is relatively the same compared to a player. It also has a strong correlation with one’s Save Percentage as well.

Conclusion

I tested these formulas on a game between the Cyclones and the Nailers, and I think it turned out very well. If someone gets 0, then they did absolutely nothing because everything that is in the formula, is all the stats in the ECHL box score. 

I do think that there is a weakness where if a goalie has a bad game, it completely shifts the data to where he has a much lower metric than the rest of the team.

But I do think it is a good measure on how to compare who had a bad game and who didn’t.