2021-22 NCAA Hockey Power Ratings
In-season power ratings adjustments are current as of Sunday, April 10, 2022.
What Is a Power Rating System?
Power ratings in sports are numbers that assign a value to each team based on a number of factors. Much like stocks, it can be advantageous to buy when a team’s value is low and sell when it’s high.
There are different types of power ratings systems, such as market-based, results-based and Bayesian. Market-based power ratings use actual betting market data to determine the relative strength of each team. Results-based power ratings rely on margin of victory and strength of schedule. Bayesian power ratings use statistical methods to compute and update ratings after obtaining new data.
All three systems have their advantages and disadvantages. Results-based ratings systems tend to do an accurate job of assessing a team’s resume — i.e., what it has accomplished to date.
Market-based and Bayesian systems tend to have more predictive power. Market-based ratings reflect highly liquid and efficient betting markets, while Bayesian systems are less prone to overreacting to (or being skewed by) a single data point than results-based power ratings systems.
While market-based power ratings systems are Bayesian in nature, they rely on betting market data, whereas Bayesian power ratings systems can be built with a variety of statistics but don’t necessarily need betting data.
NCAA Hockey Power Ratings Methodology
Our NCAA hockey power ratings system is results-based and schedule-adjusted. The ratings estimate goals for, goals against and net goal differential for each NCAA hockey team vs. a Division I-average hockey team.
We use simple linear regression techniques on widely available NCAA hockey game results. The resulting set of ratings is similar to the Simple Rating System (SRS) devised by Pro-Football-Reference.com for the NFL.
While there are many ratings systems synonymous with college hockey — such as RPI, PairWise and KRACH — we find it frustrating that most of them are represented as difficult-to-understand percentages, rather than something that all hockey fans can easily understand: goals for and goals against.
How to Use Our NCAA Hockey Power Ratings
These projections can be used as a baseline before factoring in home-ice advantage, travel, rest, motivation, the presence of the starting goalie as opposed to the backup, matchup-based adjustments, and any injury-related news.
The ratings are standardized to be on the same scale as goals — with a goal differential of zero representing an NCAA Division I-average team — so that you can simply subtract one team’s goal differential from another to get the expected goal differential on neutral ice, then add in a home-ice advantage adjustment. (Home ice was worth about 0.298 goals during the 2021-22 season.)
One of the primary goals of these results-based power ratings is to find discrepancies between public perception (based on a team’s win-loss record or recent performance) and how a team may be benefiting from factors that could be misleading, such as record in one-goal games, or circumstances beyond its control, such as schedule difficulty.