Making accurate predictions isn’t easy. Not in the NHL, not elsewhere in hockey, not in real life. The analytics communities in all major sports have taken the first steps, but we are still far from being able to predict, for example, next year’s playoff participants.
Just after the 50-game mark in the 2015-16 season, Sportsnet’s Stephen Burtch published a piece projecting the regular-season finish. He used a model that combined the best tools we currently have, but failed to make an accurate prediction for this year’s playoff teams — even though over 50 games had been played already.