Under The Hood Changes to SPAM for 2025-26

Sean’s Player Assessment Measure turns 1 tomorrow, and there are so many memories. First steps, first correctly predicted upset, the list goes on.

After 365 days of using the model and its tools, plus getting a lot of helpful feedback from users, I have a few evolutions to the model that I would like to share with you.

There are 4 changes being made to how a player’s SPAM is calculated.

The first and most important is moving Defensive Zone Exits from a defensive stat to an offensive stat.

The idea behind this change is that while exiting the defensive zone marks the end of an opponents possession and could be seen as a defensive action, it is not a measure of a player’s defensive ability, and often led to offensive minded players earning inflated dSPAM stats.

Think about the skills that go into exiting the zone on a typical play. The defenseman will either make a pass or cut up the ice towards the opponent’s zone, and a forward will need to catch a pass then make a second or carry the puck themselves. This fits much better into Play Driving, especially for high volume transition players.

The second change is that Takeaways have been moved from Puck Battling/Possession to the Offensive Disruption category. They still count on defense, but they’ve just moved over to a different neighborhood. Now that Zone Exits aren’t in odSPAM, Takeaways fit in much better and pbSPAM becomes more focused on up for grabs pucks and battling.

Third, Hits Against have been removed. I quite frankly shouldn’t have included them to begin with, but originally they were used because in InStat it looked like most hits came before a turnover or loose puck (which is bad), but realistically it just served to punish players who had the puck (and having the puck is good).

Lastly, a small tweak has been made to how Opponent Offense Against has been adjusted for Strength of Schedule. The difference is very minor, but it was one that always bugged me, so I changed it. As you can see above, it only had a 1% difference in Cole Hutson’s score for the category.

To see this formula in action, you can view this year’s player cards here. Stats are updated through January 6th, and I will be checking back in bi-weekly or more for the foreseeable future.

If you’re interested in looking at the code, I’ve updated it here, and you can refer to my original blog for any questions as to methodology.

I made these team cards just so I could do a quick whip around the top of the NCAA, as I haven’t been able to follow week-in-week-out, and I hope you’ll find them equally as interesting to look at.

They give you a quick, high-level look at who their top players have been, and where the team’s strengths lie. I’ve also included where the team would rank if you simulated match-ups against all 62 other teams, and the percentile for their Expected Wins after applying my prediction model to all these games.

Full transparency: this isn’t anything groundbreaking, but I hope that it can either entertain you, give you a rundown of a team you haven’t watched yet, or at least help win an argument online.

I’m somewhat neutral as to what the next SPAM-related project is, so I want to know, what would be most helpful to you?

The 3 ideas I currently have are below:

• The Match-Up cards would be similar to last year’s preview cards, but you would be able to plug your own teams in.

• Goalie data exists in InStat, but I would need to do some massaging to get a ready-to-use visual done beyond just GAA, Sv%, and my bodged together GSAx.

• Expanded team cards could let you see the whole roster, and maybe ice time data could be incorporated to try and show lines.

Thank you for reading and I hope that you’ve enjoyed this college hockey season so far. You’ll hear from me a few more times in the coming weeks.