What does ESPN's FPI project for Penn State following their bye week
Taking a look at Penn State Football's 2023 schedule and chances in each of their remaining games this season.
Football What does ESPN's FPI project for Penn State following their bye week
- Thread starter Dylan Callaghan-Croley
- Start date
I still think Ohio State should be in low 40...similar to Michigan's odds against us
If I recall correctly, looks like they finally adjusted Michigan State to a more realistic number. That's about as much as a lock as UMass right now
If I recall correctly, looks like they finally adjusted Michigan State to a more realistic number. That's about as much as a lock as UMass right now
So, we have a 12.2% chance of going undefeated.
What does ESPN's FPI project for Penn State following their bye week
Taking a look at Penn State Football's 2023 schedule and chances in each of their remaining games this season.pennstate.rivals.com
who even cares? Someone built a model and nobody has any clue if it is good. seems to be a lot of that around college sports. Someone builds a program that assigns a number to each team and for some reason people think this is interesting.
The OSU number is too low although I am sure the fact that we have won there twice (once when OSU had an interim coach and was in transition) in the last 30 years is factoring into their models.
Almost zero chance that would come into play.
The big reason is that they have OSU 1st and PSU 7th. Their model has Michigan lower than everyone else too (hence our inflated chances of beating Michigan).
Last edited:
In that case then I am surprised they have us at 34%. I would think it would be around 20%. Nevertheless, let's just play the game.
You know, it does bring up a valid question: how has the model performed over the years? Is there a retrospective that looks at how games were predicted right before they happened and how accurate FPI is? And if the underdog team won, what kind of chance did they have going in? A team predicted to have a 40% chance of winning and then winning is not that surprising, but if the model is wrong often when it assigns a <15% chance of winning for a team, I would question it.
doubt they do any serious evaluation. its like those in-game win probabilities that will sometimes tell you a team has 99.9% chance of winning during the game, and then they proceed to lose (i have seen that more than once)
if people click on it, that's all they care about
