topic Multiple Logistic Regression in JMP in Statistical Procedures

Multiple Logistic Regression in JMP

sblacklow — Mon, 11 Feb 2013 03:44:22 GMT

I would like to model a win or loss of the Patriots football team based on a few different predictors. Is it possible to fit a multiple logistic regression model in JMP? If so, how can I do it?

Thank you!

Re: Multiple Logistic Regression in JMP

PaigeMiller — Mon, 11 Feb 2013 14:53:34 GMT

Yes, you can do so, you would use the Fit Model command in the Analyze menu.

Speaking as someone who has followed for decades the mathematical attempts to predict sports wins and losses, my advice to you is: don't bother, there must be something better you can do with your time.

Re: Multiple Logistic Regression in JMP

SteveDenham — Tue, 12 Feb 2013 13:16:05 GMT

My first statistical program was an effort to apply Spearman's rank correlation statistic between a variety of American football statistics and final standings in the 1971 NFL (I may have just pointed out what a dinosaur I am). My instructor was impressed--the results were negligible.

Steve Denham

Re: Multiple Logistic Regression in JMP

PaigeMiller — Tue, 12 Feb 2013 13:53:42 GMT

We're in the same boat Steve

Using published statistics to predict NFL results has two serious problems:

Relevant information is not available (i.e. relative strength and speed and stamina of an offensive lineman compare to the strength and speed and stamina of a defensive lineman)
Noise swamps the signal (injuries, referee's calls, weather, coaching decisions, etc.)

As far as I know, no one has ever come up with a successful model that worked long-term (otherwise, if there was such a model, the Las Vegas bookies would go broke, and that hasn't happened)