College Football Nerds Model + Model Predictions for Some Games
Ahoy!
There's a free online prediction model offered by College Football Nerds (link). I believe it's the same thing they use in their previews. I think it's pretty handy, though you need to start an account to do more than two matchups. Here's a picture of the stuff they show on the site:
Below, I've put what the model predicts for some noteworthy games, both this week and further into the future. Winner in bold.
- Michigan 49, Indiana 10.5
- Michigan 28.1, Ohio State 31.3
- Oregon 35.5, Washington 29.8
- Notre Dame 43.9, USC 40.5
- Miami (YTM) 31.4, North Carolina 23.8
- UCLA 18.1, Oregon State 17.8
Some thoughts:
- I think Ohio State running up the score and Michigan doing the opposite impacts this prediction for The Game, but I don't know enough about the model to say. However, they mention that the model favors teams with weaker schedules, and I'm thinking running up the score has the same impact as a weak schedule.
- Not sure this takes into account home vs. away, or if that even makes a difference in outcomes. That would make me rethink USC/ND, though I think USC is essentially held together by a series of Caleb Williams miracles.
- I've seen both UCLA/Oregon State. That looks about right to me. UCLA looked really good, and Oregon State wasn't very consistent.
October 11th, 2023 at 2:50 PM ^
If you say that OSU scores more then 23 last year (but still less than 45) if they spam Marvin Harrison more instead dumbass screen passes on 3rd and short, I wouldn't have a problem with that. Massive credit to our secondary but we got a little help from Ryan Day's play calling.
I think Day is smart (AND TOUGH) enough to know he can't win by playcalling as tight this year and Harrison is definitely good for at least two catches that make us throw things at our TV. But all of that is counterweighted by the fact that McCord isn't going to find the open guy as automatically as Stroud did. We are of course going to meme on CJ for 21/22, but the guy was dropping dimes on everyone in college and is impressively doing the same in the NFL without the same talent advantage he had at OSU. McCord will obviously get better as the season goes but he won't be Stroud.
October 11th, 2023 at 3:00 PM ^
and they still cannot run the ball. against shitty competition. at least shitty defenses. They ran for 66 against maryland. They run as a team for about 100-150 on average... that isnt winning in the trenches. Also, we think their D is improved after holding ND down, but ND got 20 against Louisville and 21 against Duke, so not exactly prime time there.
I am a bit lost - and hope to be so into December, why all the trends are to pick ohio, on the road, with the same coaches, with a new QB, and a more seasoned UM team.
If i had to bet and not root, I'd predict that of the 3 teams left undefeated, PSU has the best path to going 2-0 in the 3 team gauntlet. I hope its wrong, but somehow that psu game just stinks in timing and everything else. What I have in our favor is that they'll already have had their first "huge game" with OSU.. seems like the luster wears off after that first big one. But itll be at home and psu is always nonsense at home.
October 11th, 2023 at 3:17 PM ^
That was a one time occurance though. There's a second scenario where OSU picks up the 4th down to Stover in the 2nd quarter, goes up 17-3, and never looks back.
These model simulate things 1000s of times to see how things playout. There are probably many times in this model where Marvin Harrison Jr. goes for 200 yards and OSU wins by 3.
October 11th, 2023 at 4:22 PM ^
OSU was terrible at short yardage conversions. So, no, there is no world in which they convert that.
But, is there a model where JJ hits Ronny Bell in stride 15 yards behind the secondary instead of making him stop and wait for the ball and gets tackled, leading to our first TD instead of a FG? Is there a world where Moody hits a 58 yard FG that he made against TCU and missed by inches? Is there a scenario where McGregor intercepts a pass that was all but in his hands on a screen and houses it? What about the time Stroud threw the goal line out that Sainristil had dialed up perfectly and where he would have housed an on-target throw, but Stroud overthrew?
October 12th, 2023 at 10:19 AM ^
There are a zillion "second scenarios" though. Picking that one play and changing the outcome is silly to try to prove a point. Michigan also had crticial plays that almost worked in the 1st and 2nd Q before we got rolling.
October 11th, 2023 at 1:58 PM ^
Hopefully Michigan beats OSU so bad that the nerds go back to the basement to play dungeons and dragons while listening to Rush or whatever the fuck it is that nerds should be doing nowadays.
October 11th, 2023 at 2:30 PM ^
Nerds are running the country now. That's why everything sucks.
October 11th, 2023 at 5:12 PM ^
I wish
Instead they're making $$$ creating shitty social media companies
October 11th, 2023 at 2:07 PM ^
I am mystified as to why this site gets so much play on this board. Every supposed analysis or model that comes out of it seems barely deeper than an Associated Press game summary.
October 11th, 2023 at 2:11 PM ^
The low consistency figure (it is more like the r-squared of regression) for us is the result of garbage time. OSU seems to play the same way the entire game (actually, we have one of the lowest). I am not sure how this affects the predictions, but it must.
October 11th, 2023 at 2:21 PM ^
After the last few years, I've come to the conclusion that computer models just can't completely comprehend Michigan. Our style, the lack of stat padding, players sitting out quarters of games, our run first approach, it all just seems to baffle computer models.
Anybody with eyes and two braincells should be easily able to understand why Michigan should be favored over Ohio State at this point
October 11th, 2023 at 2:29 PM ^
They have us losing to Ohio? Fuck the nerds and fuck Ohio.
October 11th, 2023 at 2:30 PM ^
Like all models, the key is validating the predictive quality against actual results. What did that model say about our previous games? Or Ohio State's previous games?
October 11th, 2023 at 2:34 PM ^
-Their model doesn't include recruiting rankings or prior year results. This leads it to getting materially different output that SP+ or FPI sometimes.
-They have acknowledged that if they had a "Magic Formula", they wouldn't share it, they would be on a beach somewhere, usurping Biff Tannen as the greatest sports bettor of all time. They were 62% ATS last week. This is not sustainable.
-The model is insanely high on Penn State - it has them beating both Ohio State and Michigan by at least a touchdown. It clearly does not have a "James Franklin *Not Elite* Clock Management" variable.
-It can be a useful tool to pick up teams that are materially different than last year by week 5-6 - the model was way higher on Oklahoma than most teams.
As an Ohio State fan, I think we beat Penn State in a game we should probably lose, and get rolled by Michigan.
October 11th, 2023 at 2:37 PM ^
It plainly does not take into account offensive line play.
October 11th, 2023 at 2:58 PM ^
I looked up their model's analysis of UM vs. PSU and it had UM losing 24-32, which is about all you need to do to determine how useful this model is.
I think the guys who make the videos and analyze the games are fine, even if I think a lot of their analysis devolves to "this team is good and this other team is good so I think the good team will win because they're good", but they've been wrong about UM a lot recently and so whatever analysis they're doing doesn't jive with reality for UM. I don't know how they are with other teams but I presume it's equally hit and miss.
October 11th, 2023 at 2:59 PM ^
Follow it for a few weeks and come back with a report on how much it worth.
October 11th, 2023 at 3:10 PM ^
MGOBLOG nerds are better.
October 11th, 2023 at 3:11 PM ^
Does this model not taking running the ball into account? Because running for 1.9 YPC against freaking Maryland, at home, when you're supposed to have this massive talent advantage at every position does not bode well for when your schedule gets tougher. Seriously, is all this thing cares about is recruiting stars and downfield passing?
Not saying OSU can't win in Ann Arbor. The minute you think they can't is the minute Michigan gets handed a big fat L to end its regular season. But OSU's "big, tough win" from a few weeks ago really doesn't look that good anymore. Or at least nowhere near as good as it did when it happened.
OSU does not look like the toughest tough guys to ever tough their way out of Toughsville like Day has been proclaiming them to be.
October 11th, 2023 at 3:27 PM ^
This model has Air Force (29.1) favored over Georgia (25.8). I'm going to go ahead and ignore this as useful in any manner whatsoever.
October 11th, 2023 at 4:49 PM ^
Their model is jacked.
October 11th, 2023 at 5:14 PM ^
This is the worst OSU team in years, and the best UM team in years.
When OSU was better, a somewhat worse UM team beat the snot out of them.
QED, the model is shit.
October 11th, 2023 at 7:06 PM ^
RE: OSU
"Y'all soft as f**k!"
That's all I need to know for Ryan Day's OSU.
Or, put another way, if you have to say you're tough, you're not.
October 11th, 2023 at 9:38 PM ^
Damn, I so badly wanted to believe in this model as I love stuff like this. But any model that has OSU beating Michigan, even if a neutral field assumption, is a pretty poor model.
October 12th, 2023 at 8:17 AM ^
Has ND scored 40 points this season yet? They are likely to get smoked by USC.
The same Miami that lost to GT? Beating UNC?
The "model" is why I don't listen to those guys lol