The first month of the NHL Model is complete and it is time to review the results!
Game by Game Prediction Results
Over the course of 226 games in the month of March, the model has accurately predicted the winner for 58% percent of the games. This percentage is right in line with the expectations I had before the month started.
It is also worth noting that home teams dominated despite the limited number of fans in attendance. This will be a trend to keep an eye on as the season progresses.
Model Results: Book vs Model Probabilities
The picks provided so far in this season have all been required to meet the 12% threshold as shown on this chart. Picks that have met that criteria have gained 6.58 Units and provided a return on investment of 109.7%. I do not expect this ROI to be sustainable during the length of a full season, but the results have been extremely promising. The 11% threshold has provided more units but at a smaller ROI. This is another thing to monitor going forward.
With the success of the model to this point, the biggest surprise has been the lack of games that have met the criteria. The last model play on this site was two weeks ago, and I expect they will be hard to come by in the coming months.
I am very happy with how the model has performed to this point and I plan on finding different ways to present picks through an analytical lens. Moving forward the model will be tracking the over/under of games as well as puck line results.
Thank you for following along and I look forward to providing winning picks and analysis for the month of April!