I don't see how trend analysis would help improve the projection to the Sabres all that much--two horrible years followed by a low-middle year is statistically a blip, not a trend. And while weighting last season heavily would certainly be appropriate for the Sabres given the tank, it would be a poor modeling choice for the league since pure luck in a given season can have such a dramatic effect on a team's record (Calgary and Colorado making the playoffs, LA and Montreal missing, and so on).
His model is fine. Sure it's not a good fit for a team that spent two years tanking, but adjusting it to account for that outlying scenario would likely worsen the model as a whole.