Tables of Stats
Registered User
Due to being a petty bastard in debating another member of the Canucks section of this site, I've done a stat projection for your team's upcoming season. It's a simple model that averages the stats from a player's past 3 NHL seasons. It's more fully described as:
1) For each player take the average of their last 3 seasons for games played, goals, assists, and points. I used to use AHL data for games played to measure player durability, but I have used only NHL data this season.
2) Fill out the team's predicted roster using 14 forwards and 7 defensemen.
3) Compare the predicted number of man games played to the number of man games played over a full season for both forwards and defence.
4) Fill out the remaining games played with the average stats of the bottom of the roster players.
This model tends to underpredict points for younger players who are still improving and over-predict for older players who may be declining.
The depth sections are used by scaling the man games available to the model with the stats produced by your 15th and lower-rated forwards by PPG. This means that players promoted up the depth chart may be counted twice, but this error should be small enough that it shouldn't impact the overall prediction.
Kane is predicted but his stats aren't included in any totals as his ability to play this season is still unknown as is his expected impact if he does return.
I hope you all find these predictions as interesting as I do.
1) For each player take the average of their last 3 seasons for games played, goals, assists, and points. I used to use AHL data for games played to measure player durability, but I have used only NHL data this season.
2) Fill out the team's predicted roster using 14 forwards and 7 defensemen.
3) Compare the predicted number of man games played to the number of man games played over a full season for both forwards and defence.
4) Fill out the remaining games played with the average stats of the bottom of the roster players.
This model tends to underpredict points for younger players who are still improving and over-predict for older players who may be declining.
The depth sections are used by scaling the man games available to the model with the stats produced by your 15th and lower-rated forwards by PPG. This means that players promoted up the depth chart may be counted twice, but this error should be small enough that it shouldn't impact the overall prediction.
Kane is predicted but his stats aren't included in any totals as his ability to play this season is still unknown as is his expected impact if he does return.
Code:
https://docs.google.com/spreadsheets/d/1GUv5Wo01jvZ14_03udIfErlGSzhXJS7oGvzJlyLlSq4/edit?gid=0#gid=0
I hope you all find these predictions as interesting as I do.
Last edited: