The Nanooks welcomed the Huskies to the Carlson Center in Fairbanks this weekend and Tech was able to come home with two regulation wins. That’s something we all love to see, but something else I love to see is numbers and charts that will hopefully help me better understand how the team played this weekend. If that’s something you’re looking for then don’t change that dial because you’re in the right spot.
Thursday October 13th: Michigan Tech 2 – Alaska Fairbanks 0
Despite being the closer game on the scoreboard, I actually feel like this overall was the better game we played this weekend. As you can see in the graph below, the boys came out a little slow, but after the first goal was score near the beginning of the second period they really took control of the game and didn’t let go.
Another fun was of seeing how the game developed over time is a new addition to the shot maps that I have been making for the preview articles. Through a little coding tomfoolery I was able to create an animation that populates the shot map over time.
Aside from the difference in quantity of shots, something I think is very noticeable from the shot maps is how Tech was able to take shots from much more dangerous locations. One of the first things I noticed looking at this was how at the beginning of the game there were a lot of shots from both teams from the point and from outside the faceoff circles. Around the time Tech scored their first goal and the rest of the second period they were driving the center of the ice and getting a lot of shots from the slot. Which, unsurprisingly is where they scored both their goals
On the other side of the ice, about half of Alaskas shots were taken from the blue line. A shot from the blue line is almost never going to beat a goalie in the NCAA unless he is screened or it is deflected, and while it doesn’t show in these maps from what I remember in the game Alaska was not doing a good job of screening Blake and getting in position to defect or get rebounds. Overall the team did a good job limiting dangerous chances for Alaska and so I am not at all surprised that Pietila lock it down and add another shutout to his collection
This is one of my favorite charts, this chart is what +/- stays up at night wishing it could be. The X-axis is meaningless, its just spread across the X-axis for readability. What matters here is the Y-axis, which shows for each player while they were on the ice, what percentage of all the xG produced by both teams was produced by that players team. So for example, in this game Logan Pietila was on the ice for 1.26 xG for Tech, and 0.66 xG for Alaska. The formula to calculate percentage share is this (xGF/(xGF+xGA))*100. So you plug in your numbers and (1.26/(1.26+0.66))*100 = 65.6% which if you look at the chart that is exactly where Logan is on the Y-axis. Pretty cool stuff.
Friday October 14th: Michigan Tech 6 – Alaska Fairbanks 2
Game two despite actually looks worse according to most of the xG metrics which is a little surprising given the actual score. My theory is that Tech was able to get a big lead off of a tough night in net for Radomsky, and from about the halfway point of the game they shut it down and tried to focus on holding the lead instead of generating opportunities on the attack. Those of you who read my intro to advanced stats article earlier this week might recognize this as a place to apply Score Effects. And yes, that would likely make the charts look better for Tech. I didn’t do apply score effects for two reasons in this case. First because its most important to calculate that out when looking at a large sample size like an entire season. And second, because there hasn’t been much research done into how significant score effects are in NCAA hockey and I haven’t had the time to do my own research on it to make accurate calculations.
Something you can see from the shot map is this was a game of quantity vs quality. Alaska once again took a lot of shots from the blue line, which aren’t very dangerous but they did take a bunch of them which do add up in the xG charts. Meanwhile Tech took barely any shots from the point or near the boards.
Okay so I have a confession to make, in the last section talking about this chart I said a lot of nice things and didn’t mention any flaws. The big flaw for this chart however is it doesn’t account for time on ice, so you can absolutely get outliers who look incredibly good or incredibly bad. In the first game there was pretty good distribution of ice time and I didn’t immediately notice any outliers. This game has a few though, for example Erriks Zohovs for UAF is near the top of the chart which should mean he had a great game. But in reality he only played nine minutes and just happened to have one good shift. On the other end of things, Marcus Pedersen is way down at the bottom, but he only got on the ice for six minutes and likely had one bad shift that tanked his on ice xG stats for the night.
There are ways to account for the amount of ice time to try to determine how well a player played during his time on ice, but they all rely on a larger sample size of games than the three Tech has played so far. But it’s something to look forward to being able to add into charts like this in the future.
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.