CUOrange Blog

Clemson Tiger Sports…and other stuff…

  • Blog Stats

    • 11,268 hits
  • You don’t ever know about a football coach until you get in a ditch. He worked his way out of a ditch pretty good. That’s a good sign. - Danny Ford on Dabo Swinney
  • Recommended: AE’s Tiger Gear

    Adam is a Clemson grad, fellow CUTiger.com member and an all around good guy. Click here to check out all of his stylish Clemson related products.
  • RSS Clemson Basketball

    • An error has occurred; the feed is probably down. Try again later.
  • RSS Clemson Football

    • An error has occurred; the feed is probably down. Try again later.
  • Mystery Number

    3,496

Computer Model Forecasting

Posted by cuorange on May 29, 2009

As I was putting together my Regional Predictions earlier in the week I got to thinking about the shortcomings in my selection process.  The biggest being that I didn’t have a clear way to weight the factors that I consider important to winning the national championship, namely fielding, pitching, batting and slugging (in that order).  I looked at the numbers, put more weight on the ones I believed were important and out came a pick.

This oversight and a natural tendency to favor one team over another (example: Hitler over South Carolina) made me question the model.  Add to that the fact that the model didn’t include a strength of schedule component and I thought there had to be a better way.

The strength of schedule is important because in determining the better of two teams with similar stats it’s obviously important to compare what teams the stats were compiled against.  For example, Bethune-Cookman’s 41st ranked pitching sounds pretty good, but it was earned against the 214th ranked schedule.  Compare that to Oklahoma State’s 42nd ranked pitching, which was earned vs. the 11thtoughest schedule.  In the “human” model only raw numbers were considered, placing B-C slightly higher than OSU.  In the real world I think we can agree that OSU has better pitching than Bethune-Cookman.

I also went back and looked at the last 7 national champions to determine weights to allocate for these 5 factors and went to work developing a computer model.  This model heavily weights fielding and the strength of schedule as those are the traits that most highly correspond to winning championships, followed closely by pitching.  Much lower in the pecking order is batting average and slugging percentage.

This model is still not perfect in that it doesn’t include a home field advantage, but using this model brings us closer and reduces the human element.

Here are the regional picks using the computer model, with the seeds in parentheses and my “human” pick (where different).  We’ll compare the two results next week, but there are some interesting results here:

  1. Texas (1)
  2. Texas A&M (2) Human Pick – TCU (1)
  3. Georgia Tech (1)
  4. Florida (1)
  5. Oklahoma State (3) Human Pick – Alabama (2)
  6. Arizona State (1)
  7. East Carolina (1)
  8. North Carolina (1) Human Pick – Coastal Carolina (2)
  9. Cal State – Fullerton (1)
  10. Vanderbilt (3) Human Pick – Louisville (1)
  11. Georgia (2) Human Pick – Florida State (1)
  12. Oklahoma (1)
  13. UC-Irvine (1)
  14. Mississippi (1)
  15. Kansas State (2)
  16. Louisiana State (1)
  • The first thing that pops out is the Oklahoma State and Vanderbilt as 3 seeds. 
  •  The worst number 1 seed is Florida State and the Tallahassee Regional is also the weakest in the field. 
  • The worst team in the field is Southern University, followed closely by Binghamton.
  •  We both agree on Kansas State over Rice. 
  • We both agree that Fullerton is the #1 seed. 

It’s important to remember that each team’s numbers will change as the playoffs continue and the rankings shift, so while we might predict a Texas win over Texas A&M in the Super Regional as of today, that might not be the case next week.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

 
%d bloggers like this: