Was The Algorithm Right About Alabama?

When I first posted my rankings and tried to get some attention on them  via Twitter (@rankingsray), the Bama fans lost their minds over how low I had them, and more importantly that I had other "less worthy" teams above them.

You can go check out the earlier posts (please do!).

My first ranking was released Nov 17, and the CFP released their 3rd rankings Nov 19. Please keep in mind that I was still refining the system (and I still am), so there is probably more movement than we would like, but we can still make the point.

CFP Week 3 Alabama rankings:

CFP #5
AP #5
RR #16

You can probably see why there was a little pushback.

CFP Week 4:

CFP #5
AP #5
RR #17 they dropped because Bama played Western Carolina that week.


CFP week 5: Tide loses to Auburn:

CFP #12 (-7)
AP #9 (-4)
RR #14 (+3) I added the SRS component which helped

CFP week 6: Now

CFP #13
AP #10
RR #14

My system started at 16 and ended at 14, while they both started at 5 and ended 5 (AP) and 8 (CFP) spots lower.

I am not trying to bash Alabama here. As I've said before, I simply want to to provide an unbiased system for ranking college football teams that don't play each other except for at the end of the season (and I would really like an expanded playoff). Alabama is just convenient this year. I haven't set up the spreadsheet yet that will allow me to do a more thorough and methodical examination of possible bias. If you've read my look back posts, you can see that some years (so far) have been very similar between human polls and this system, but that is only end of regular season polls. I plan on building out my spreadsheet so that it can track changes over the course of a season. Or I may have to teach myself something else to crunch this data, maybe R or something Python based. I'm no coder, but I can try!

Remember, this system is focused on "what have you done" not "what do I think you'll do." The question is, what are the AP and CFP polls trying to do: predict or measure?