Help me come up with a Project Design

by mgrex03 on Mar 5, 2008 10:22 AM EST in Analysis Comment 8 comments

Large Forehead = 80% Win

In the same procrastinating fashion I used during my undergrad days, I have a project proposal due tomorrow that I have not started, and I need your help. The class is on statistical modeling, and I have to come up with a model of my choice to present to the class. Naturally, with my new position here at Stampede Blue, I want to create a tool that predicts NFL games.

This tool will be a decision based model, meaning all the factors involved have a certain probability of happening, which causes other things to happen, etc. Here's where you come in:

I need ideas as to what factors go into predicting a football game. I think I have the statistical parts of it, but I need non-statistical factors as well, such as weather or size of quarterback's forehead (I'm not really using that one).

As a thank you, I'll post my report, as well as the tool, here so everyone can check it out.

0 recs | Comment 8 comments

Story-email Email Printer Print

Comments

hm

specific stadiums (homefield advantage)
starting quarterback injured for a team
monday night games/thursday night games

by yellowsnow on Mar 5, 2008 11:29 AM EST reply actions 0 recs

taking a shot

previous weeks result would be an intersting study. i.e. - previous travel or short week

time of game would be interesting as well, does a certain team perform better during an early game (noon) or late game (3) or even primetime games.

www.behindthesteelcurtain.com

by cgolden on Mar 5, 2008 11:45 AM EST reply actions 0 recs

or even

the previous week being a hyped game against a tough opponent, brining in the possibility of a let down, or trap game. (You may have to factor in that more experienced teams such as the Colts or Pats don't seem to be as vulnerable to those traps though.)

The Shogun of Harlem

by shonuff on Mar 5, 2008 12:17 PM EST up reply actions 0 recs

starters

How about the number of starts each of the starters has had for that season? This might factor in injuries and experience for your players and how consistently a group of guys is playing together (was our revolving door of an offensive line that made me think of this one)

or how about if this is a divisional, conference or non-conference game?

by bluegirl on Mar 5, 2008 12:24 PM EST reply actions 0 recs

Market Size?

I'd be interested in seeing an analysis with some revenue figures thrown in exploring winning % over several seasons as a function of market size or some other economic factor.

by Zumone on Mar 5, 2008 12:37 PM EST reply actions 0 recs

let vegas do the work for you

a stat on how teams teams historically do against the spread (since vegas figures in a lot of those other factors anyway)could be informative. you could factor in the size of the spread and whether its a win loss or cover...

Tom Brady can throw a football pretty far. Bob Sanders can throw Tom Brady even farther.

by madjediskills on Mar 5, 2008 1:00 PM EST reply actions 0 recs