The win and cover probabilities below are based on data from the 2013 and 2014 seasons and will change as games are played in 2015. Think of it as a baseline number that will be adjusted when additional information becomes available.
Some will question these numbers, specifically Louisville and Florida State. That’s fair. I’ll reiterate that these are intended as a starting point and not a “final answer”.
These probabilities look at projected metrics, compare them to the last 2,841 games between Division 1-A teams and compute how often a team with those projected metrics win (and cover the spread).
But still, why is Louisville rated so highly? In a word – defense. The Cardinals led the nation in 2013 and weren’t far behind in 2014. Add to that a high turnover margin you have the makings of a victory.
Many, including me, thought Clemson would roll a Teddy Bridgewater-less Louisville at Death Valley last season, but the Tigers needed a fourth down stop to hang on.
Another question is how did I factor Deshaun Watson into the Clemson numbers? The answer is that he figured just as much as he has contributed. It’s tempting to assume Watson will play 12 games and put up video game numbers, but history shows that’s not likely to happen. As the 2015 season unfolds and (if) Watson stays healthy he will have an increasing impact on projections like these.
It’s an imperfect science for sure and harder to decipher and put into projections such as these is how do you value someone like Grady Jarrett? What does Jarrett mean to the Clemson defense in terms of yards and points given up?
There are two other tendencies to mention, recency bias and the “my team is getting better and everyone else is staying the same or getting worse” bias.
Recency bias is typically a financial term, but used here it means remembering what happened most recently (crushing Oklahoma and romping vs. South Carolina), and forgetting the general struggles of the 2014 season.
The “my team is getting better” bias is a little more nebulous, but generally happens when a fan lists all the reasons why his or her team will be better, then mentions all the talent the opponent is losing without mentioning any of the talent (or coaching, etc.) improvements of their opponents.
Take these for what they are worth at this point – a work in progress and something to talk about during the offseason.
My projections bested the ESPN FPI numbers in a small sample size (ACC Bowl games) last season, so I’m fairly confident that, as a whole, these numbers (or more accurately the numbers I post next fall) aren’t way out of line.