AJHL Final Year Performance as a Predictor of NCAA First Year Scoring

Turns out final AJHL year stats is not a great predictor of first NCAA year.

This year I’ll be attempting to cover Arizona State University’s NCAA hockey program as part of the Hockey Graphs Mentorship Program. One of the projects I’d like to create is a point projection tool for the freshman, so this was a first pass as a practice.


To build a sample size of over 50, I referenced Elite Prospect’s rosters for each NCAA team. I found players who had played in the AJHL the year before going to the NCAA, and played in a few games their freshman year. I plotted the relationship between the points-per-game of the final AJHL year and first NCAA year in an attempt to find a relationship that could be used to create a projection.


At least with the data I collected, there is not a good relationship between the two seasons’ points-per-game.

An example of a not-good fit.

The highest R squared value I could achieve without bizarre functions was 0.21 with a second order polynomial. 79% of the success freshman year for AJHL alumni is not explained by the AJHL final year performance. Some other variables that should be investigated (if I have time to go back through this, data collection took about 3 hours) are: height, weight, change in points-per-game throughout the AJHL career, and quality of teammates.

Based on the model, Cale Makar (freshman at UMass Amherst, AJHL alum) is predicted to get somewhere between 3 and 26 points. When I tweeted this projection, I received some positive feedback.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.