![Old](images/statusicon/post_old.gif)
4th September 2012, 09:57 PM
|
Member
|
|
Join Date: Sep 2011
Posts: 1,492
|
|
Well I ended up writing a basic extract to get some of the team data. I would be interested in some help on my process.
I m thinking I should do some regression analysis of each variable against the points scored and points conceeded and points difference.
My aim is to find the variables that matter. I also have some not so obvious variables which are derived from the stats which look promising.
After I determine which factors in each game have a good coefficient I am then going to get the player data.
I am then going to do some trend analysis on player performance to see if I can come up with any patterns. Hopefully from this I can predict a players contribution to a game for each variable that matters.
If I get to this point and its "reasonable" at prediction I should then be able to analyse the team list for each game and predict the values which matter to winning a game based on the players who will play in the game. From this I can determine a price and whether or not I am getting value.
Does this sound like a good plan for analysis ? I have given myself at least 6 months before the season starts again to get all of the above done. Am I missing something crucial ?
|