Every modeling program I have written did very well save for thursday Frisco game. All 4, including the SPSS model that has srved me so well, failed on the Ravens 9 ers game. I am completely devoid of any emotional interest in each team, so I went in with no bias Every model showed 9 ers manhandling Ravens. I am wondering if now it is time to introduce a distance traveled unit. It is the one underlying factor I ignore due to fact this isn't the train era, NFL teams use planes, and distance is irrelevant.
I am running tests on several software variations I use to see if distance and the time between games makes a difference, and then distance by itself. IS this a variable any of you use ,and if you do is it afforded anything more than ancillary weight?
I want to believe it was a fluke. My models show that stuff, as it were, happens. Nevertheless, I am trying to flatten the days between game, distance traveled to next game, and every variation of same. Anyone else run into this, and am I ignoring instinct (knew 9ers would win) to the point of beating my own software.