Who got the NCHC right?


Well that was fun!

Now that the season is over, let’s take a quick look back at all the preseason predictions and see who was the most accurate:

NCHC Writers Poll SBN College Hockey USCHO – Candace USCHO – Matthew Metric-based Model Actual Finish
1 North Dakota Miami Miami Miami North Dakota North Dakota
2 Miami North Dakota North Dakota North Dakota Omaha Miami
3 St. Cloud State Minnesota-Duluth St. Cloud State St. Cloud State Minnesota-Duluth Omaha
4 Denver Denver Minnesota-Duluth Minnesota-Duluth Denver Denver
5 Minnesota-Duluth St. Cloud State Denver Omaha St. Cloud State Minnesota-Duluth
6 Omaha Western Michigan Omaha Denver Miami St. Cloud State
7 Western Michigan Omaha Western Michigan Western Michigan Western Michigan Western Michigan
8 Colorado College Colorado College Colorado College Colorado College Colorado College Colorado College

NCHC Writer’s Poll 2015

SBNation College Hockey NCHC Previews 2015

USCHO NCHC Preview 2015

My metric-based model, preseason 2015

As it turns out – nobody really. The NCHC Writers, SBNation and my model all had 3 of the 4 home ice teams. The writers and I also had North Dakota winning, but I was the only one who didn’t have North Dakota and Miami at #1 and #2. However, my analysis saw¬†UNO’s success coming. Wish the statistics would have seen Miami, too.

So, I’ll give each prediction 1 point for every place each team is off from the final (e.g., my Miami-at-6th prediction would incur 4 points, since they finished 2nd. My UNO-at-2nd incurs 1 point, because they finished 3rd. No points incurred for correctly picking North Dakota to win). The predictions with the lowest total score would be “most accurate.” Therefore, in order of accuracy:

  1. NCHC Writers (6 points)
  2. Metric-Based Model (Me) (8 points)
  3. USCHO – Candace (10 points)
  4. USCHO – Matthew (10 points)
  5. SBN College Hockey (10 points)

So, there you go. Turns out nobody did particularly well. But that’s what makes the season fun. I’ll write more later about tweaking the predictive model. Another full season of data will improve the accuracy. For now, on to the postseason!

Leave a Reply

Your email address will not be published. Required fields are marked *