Introducing TOEWS

Do you wish that there could be more statistical analysis in college hockey? Chances are, you just said “hell no” but in case you are intrigued, I am going to be introducing a new ranking system this year where I rank teams solely by their statistical output (though, not goal difference because goals are pretty lucky and fluky events sometimes) and see how they compare with the various national polls and the PWR.

I’m not going to claim that it will be the most accurate thing out there, especially because this is just year one and I have some kinks to iron out (mainly, adjusting for strength of schedule. That’s going to be tough to do until I get more data in there), but I have assembled a database (aka, an Excel spreadsheet because I can’t write code) that is going to help rank the teams from 1-58. I’ll probably unveil it in a couple months when it gets enough data to be reasonably accurate, and it will run alongside my other rankings. My purpose for doing this is to basically see areas where I (or the pollsters in general) have any biases and then examine why.

As for the name, it is taking in the vein of Nate Silver’s PECOTA, which projects baseball player (and team) performance. It stands for Player Empirical Comparison and Optimization Test Algorithm and is named after former major leaguer Bill Pecota. Hence, TOEWS, or TOtal Empirical Weighted Shots. Maybe a bit of a dumb name but I thought of it in about 10 seconds and it stuck. So there.