We absolutely destroyed the bye week to move up a spot in the rankings, this weekend against Lake State should be a walk in the park right?

This chart is a joke and if you are reading this I hope you have a lovely day

Friday, December 16th Michigan Tech 0 – Lake Superior State 3

This is some hockey gods idea of a joke isn’t it? I angered the gods last week by reveling in the insanity that was Ferris sweeping Mankato in a series that the numbers say they had no right to even hope for one tie. The hockey gods were swift with their punishment for my hubris. Tech didn’t look incredible, but they absolutely looked like they were the better team and just couldn’t get a bounce, especially in the third period. These are the games that hurt the most to watch, but I try to take some comfort in seeing the good things the team did, even if they didn’t get the results this time.

MTU 5v5 xG in black and LSSU 5v5 xG in yellow mapped over time. Team logos indicate a goal being scored

I don’t love seeing this many shots from the outside, but even if you exclude those the Huskies still got the puck to dangerous areas more than the Lakers. If you were one of the poor unfortunate souls who joined us in the THG gameday discord thread on Friday, you might have seen this already. But for everyone else, there was a conversation about how the Huskies were shooting everything “center mass” of the goalie. I don’t have any numbers to confirm or deny that, but from watching the game I also had that feeling. Fellow THG writer, JZ suggested the idea of scaling xG based on where on net the shot goes. Which at first glance sounds like a great idea, but I wanted to take a minute to explain why I personally wouldn’t want to do that.

The reason I wouldn’t include shot placement is that would change what xG is attempting to show as a stat. The aspects of play that lead to getting shots from specific areas of the ice are more consistent and predictable. Where on net the shooter puts the puck is much less predictable. The point (in my opinion) of xG isn’t necessarily to say if a team “deserved” to win a specific game or not. It’s to see if the team did things that will more often than not lead to success. With that mindset, I would care much more about the factors that are likely to be predictive of future outcomes. Which is (at least with the data that is currently tracked) where on the ice the player is when they shoot the puck.

Once again Brett Thorne leads the game in ice time and its not really that close. Nearly every Tech player outperformed every Lakers player with only Willets cracking the 50% mark for LSSU. You don’t expect to lose many games like this. This chart feels a little deceptive to me, because of how much Tech controlled the xG% it feels like Works, Crespi, Saretsky, and Caderoth were right around 50%. But that is actually between 60-70%. We can thank B. Huggins for being the outlier that breaks the scale on this one.

The Y-axis is the percentage of expected goals produced by the team while each individual player was on the ice. X-axis is time on ice.

Saturday, December 17th Michigan Tech 5 – Lake Superior State 1

This is what was supposed to happen on Friday. Thankfully, the gods deemed Friday’s game punishment enough and Saturday was a pretty normal game between a ranked team vs a team who has three wins at Christmas. It was cool to see Max Vayrynen get a start and good for Blake to get some well earned rest. I thought Vayrynen looked good, although he wasn’t tested very much

MTU 5v5 xG in black and LSSU 5v5 xG in yellow mapped over time. Team logos indicate a goal being scored

Holy wah, Lake State got absolutely nothing going offensively on Saturday night. Tech came out to play from puck drop. Worth noting, for some reason the only thing that was tracked in InStat for the first period of this game after 3:12 was goals. Not sure what is up with that, but their video only has the first 3:12 minutes of the first period. All of the second and third period are there. That is why LSSU has literally no xG in the first and Tech has a spike early on and then just steps up with each goal. Again, the first period is messed up in the data so it starts out slow. But Tech still did an excellent job limiting chances for the Lakers in this game.

Jack Works believers rejoice, he got 19:19 minutes of ice time which is a little over 4 minutes more than his previous high of 14:52 against BG. Tech absolutely dominated play while he was on the ice, his fancy numbers look good and he had a goal and two assists. That’s whatcha like to see. Brett Thorne with another casual game of 30 minutes of ice time.

The Y-axis is the percentage of expected goals produced by the team while each individual player was on the ice. X-axis is time on ice.

No thoughts, just vibes

A lot of what I do for these articles is data visualization, this section is vibe visualization.

The game Saturday was good, but it was really just what I expected to see both nights so seeing it happen on Saturday just kinda reminded me of how frustrating it was on Friday night.


Two weeks in a row with bonus content?? What can I say, its the holidays and I’m in the giving spirit.

This is a fun new chart I’ve been working on based off the stat Goals saved above expected. This is a pretty easy stat to understand, and I think it’s a pretty good way of analyzing the witchcraft that is goaltending without selling your soul or something. [1] The was it is calculated is by taking all of the shots taken against a goalie, calculating the xG for each of those shots and adding them all together and then subtracting the total number of actual goals the goalie has allowed. For example if I was to look at a completely randomly chosen goalie, like say Blake Pietila, I would see that he has a Goals saved above expected of 15.01 and that puts him at 5th in the NCAA. The calculation for that ends up being:

Expected Goals – Goals Against = Goals Saved Above Expected

46.01 – 31 = 15.01

This is obviously not a definitive ranking of how good a goalie is, but its one that I really like. You can also run the calculation with time on ice included, to try to account for how much the goalie is playing. But I don’t find that to be as important when it comes to goalies as it does with individual skaters.

[1] I don’t know how witchcraft works, I am not a goalie

Closing Time

I hope you enjoyed taking a look at some nice charts and graphs with me. If you have any questions about the charts, my analysis, or just want to say hello, please leave a comment or reach out to me on the THG discord server (@Augie) and I’d be happy to try and answer your questions.

Special thanks to Livonia Technical Services for being the official sponsor of all InStat analytics at Tech Hockey Guide


  1. Hey Augie, it would be great to see that goalie comparison for all four teams in the GLI. I’d be interested to see who has the analytical advantage in the tournament.

Comments are closed.