fajitapete
Six-man fan
coachGtca":q9t1fky5 said:
Very nicely stated.
of forfeits and Athens
- Thread starter granger
- Start date
Very nicely stated.
allen":12koznb1 said:
Really? Good grief
justobserving
11-man fan
O.K. Granger - I want to see the "algorithm" for this scenario....
fajitapete
Six-man fan
justobserving":3gblhvlt said:
Well it's not that pretty of an equation. Most likely to use a cross-validation statistical method for validating the predictive model. WHat you do is hold our subsets of the data for use as validating sets; a model is then fit to the remaining data and used to predict for the validation set.
You then average the quality of the predictions across the validation to obtain a an overall measure of your prediction accuracy. Best way to do this is to leave out a single observation at a time; (the game in question).
This helps to avoid self-influencing the results. Doing a cross validation on the normal linear regression helps to predict a y value for each observation without using that observation.
I don't know how many predictor variables Granger uses in his regression model but adding predictors will reduce the residual sum of squares. By doing a cross-validated we can see the error decrease if the game has value and the error increase if it is a worthless predictors.
here is an example of the equation:
View attachment 6man1.jpg
just saying . . .
Honestly, the way this is folding out, the game should just be eliminated. Sounds as if that may be the best result for everyone.
cowboy91
11-man fan
allen":bj88vbmn said:
Don't do this allen. Your schools name has been drug through the mud enough. Sounds like the administration may be trying to correct and fix things. They don't need you doing this on a public forum viewed by the entire state. Just stop. Completely my opionion though. You start slinging mud on here I promise you will eat a little, so don't throw that out there or it will br thrown right back at you.
