The action has been fast and furious on the free agent market, and yet there are still some players left that could have a serious impact on my results below. I addressed some issues with the bottom feeders having win percentages that were way too low. In the process it seems to have flattened out the standings, but I think I am a lot closer to the results I am looking for. I still have a problem predicting what a player like Steve Stamkos is going to do. He had a good second half, but his first half is almost certainly lower than what he will produce this year. Due to the lack of NHL data on Stamkos my algorithm has some problems predicting his stats accurately. I'll work on that for the next revision and hopefully by then some other sites will have some predicitons for players like him that I can utilize.

I must reiterate that this is based solely on statistical analysis and predictive models. My personal feelings as to how teams will finish doesn't play into it and therefore you can't say "You have team A way too high/low". While I might agree with you I can't argue with the numbers, especially considering how accurate they were in predicting the NHL Playoffs last season. For example in this result set I question the Flames and Flyers position, but looking at what the model predicted I can understand why it came up with what it did. Calgary adding Bouwmeester on the back end is a great signing, but losing cammalleri's goals up front makes them strapped for scoring. Similarly with Philly, they added Pronger but lost Lupul and Knuble. Obviously there is potential someone will step up and fill in those holes, but I think the model did a good job of identifying those players (Moss and Bourque in Calgary, Giroux and Briere in Philly). Therefore I can't even recommend an adjustment. I will likely do two more updates prior to the season starting and will continue to adjust as the season is in progress.

Revision 2 of this article is available here