FWIW, HockeyViz pegged us at 87 points last year and correctly projected 13 of the 16 playoff teams (they had New Jersey, Minnesota, and Calgary in with Vancouver, Washington, and Nashville out). They correctly predicted the worst 6 teams in hockey, although they had them in a different order. Like most models, there are some individual things that look absurd as a single data point, but last year's projections were decent vs the results of the season.
Their D valuation looks as bad as every other public model's does. Perunovich has the most positive impact of all our D and Gudas, Severson, and Gustafson are all top 25 D men in the league. I still haven't seen a model that is both good at evaluating D and predicting team results.
All in all, their model is projecting a big improvement from the bottom of the league. Last year they had 7 teams with an average point total below 83, with San Jose and Chicago tied for worst at 76. This year's just has us and Chicago below 83 tied for last at 81. While I'm not expecting us to be a bottom 2 team in the league, I do agree that the truly terrible teams from last year should be much closer to competency this season.