Hi again, folks!
I thought about this for a little while more which only strengthened my opinion that my system does not allow me to realistically predict anything. This is fine, as it is how it's designed. Of course, this kind of ruins the fun of making a list now and then seeing how well we do at the end of the season.
I did say in my last post that I'd have something to play with for a preseason Top 25. Here's what I came up with:
- Take all of last season's rankings: mine, AP, CFP and add up their totals. For example, Bama was #1 across the board so their current total would be 3.
- Add in 1/2 of the rank using ESPN's SP+. As I mentioned before, SP+ attempts to account for player and coaching changes as well as carryover talent. Why only half? Because that doubles the power. See #3
- Add the results from #1 and #2 to #3 and the lowest score is the preseason #1
- If a team was not ranked in the top 25 in any poll, it was awarded 26 points per poll missed (13 for the SP+).
See? Not too complicated, no opinions or votes. Of course, the AP and CFP have their biases, but you know my system is straight math and the SP+ is pretty close to that as well, so it's a wash. And I did give more power to the predictive SP+.
There are some shenanigans going on here, though.
I'm not sure I'm buying Cincy and NC, but I could be wrong there. Definitely not buying Coastal Carolina or BYU! That said, I like Liberty's chances. Notre Dame is riding the end of the season CFP love I think. Anyway, I'll keep this list in my spreadsheet and see how the season turns out.
What do you guys think?
I'll have the finalized ranking formula in a few days. Last season's sheet got so hammered due to missed games that it can't be used now, and 2019 was missing some pieces. I'm still struggling to find a way to reward a good win and penalize a bad loss. I know that's basically what SRS does, but the SRS only looks at scores, not ranking differentials. I'm working on it!
No comments:
Post a Comment