I started developing my own model around 1992 during Bama’s championship season, inspired by what Sagarin was doing and thinking I could do it better. And I did - in my opinion, Sagarin’s ratings are some of the worst algorithm-derived rankings out there. While the whole nation, including Sagarin, was saying Bama didn’t stand a chance against Miami, my system was saying Bama held a slight edge. We all know the outcome (Roll Tide!).
At that time, I was hand-crunching all the numbers for what was then around 120 division 1A teams, with pencil and paper and a calculator every week. This was incredibly time-consuming, but also very gratifying. During the 90s, I continued fine-tuning my algorithm and eventually, with the maturation of home computing technology, migrated everything over to excel spreadsheet. But it still required manual data-entry on a weekly basis and was still a full-time hobby. Around 2002, I developed my own website and contacted Kenneth Massey, who published (and still does) a compilation of all the well-known computer rankings, including his own at masseyratings.com. For awhile he included my rankings in that compilation.
Mind you, that decade of development and publication (1992-2002) coincided with a decade of relative free time for me. Then came marital engagement, a family, a son, and well, life. I just couldn’t keep up with the time required for the hobby anymore. What I needed to continue modeling was to develop a way to automate the data ingestion and move my model from excel to something else - a true databasing platform, so that there would be little to no manual time necessary from me to keep it going. But I didn’t really have the time or know how to make that happen.
One of the last triumphs of my model before I quit doing it altogether was in using it to win one of the early Tidefans pickem contests in the late 2000s. Who knows - had I been able to automate everything properly on a databasing program, maybe mine might’ve eventually been included as one of the BCS models.
Last edited:
