Welcome to our daily college football newsletter. If you like what you read and want it delivered to your email inbox every morning, subscribe to our Substack.
**
A word of warning: We’re going to get pretty wonky here about our model. I promise, we have a point.
Our college football model is primarily built of two parts: First, there’s the simulator and playoff ranking calculator which helps us forecast how likely different teams are to make the College Football Playoff. Second, there’s Movelor, our rating system. Movelor’s purpose is to predict college football games with as much accuracy as possible. It’s not the best in the world at this—there are better models out there, and nothing publicly available is as accurate as the closing spread in betting markets—but it does a solid job, and it gets stronger as time goes on. Year over year, we add improvements. Week over week, Movelor improves itself.
This last part is what we’re going to talk about today. Movelor getting better every week. Why do we think you’ll care? Because it’s going to tell us which teams still might be overestimated and underestimated nationally.
The way Movelor gets better, week over week, is by looking at final scores. It predicts the score of a game, sees how wrong it was, and then adjusts its ratings so it won’t be as wrong the next time. This is very simple. You could do it on your own, if you wanted to take the time. Take a notebook, write down how many points better you think different SEC teams are than one another, and adjust your ratings every week depending on how the games turned out. Movelor’s a little fancier than this—it uses logarithmic functions, which aren’t particularly fancy in this realm but do come from the portion of math class most of you probably tuned out—but this is the gist of what it and even the forces driving betting markets do. Miss a final score by four points? Don’t change your ratings much. Miss a final score by 64 points? You should change your ratings.
Eventually, through this process, water finds its level. Movelor is never perfect, but midway through the season it starts getting an accurate enough read on each team that the team generally overperforms Movelor’s expectation half the time and underperforms it the other half. For some teams, this takes longer than it does for others.
That’s where we’re going with this. Who keeps overperforming Movelor’s expectations? Who keeps underperforming them? These are, in one sense, the teams Movelor was the most wrong about in the preseason. It’s more believable that Movelor is still wrong about them than it is that Movelor’s wrong about, say, Northern Illinois, whose rating is back to exactly where it started the season.
We’re going to look at twenty teams here—eleven repeated overperformers and nine repeated underperformers. They aren’t the only teams who are undefeated or undefeating against Movelor’s spread, but they’re the most noteworthy on the list. We’re going to look back at their games, check the circumstance, check how badly Movelor’s missed, and make a guess on whether Movelor’s still overvaluing them, still undervaluing them, or properly valuing them now whether it missed earlier in the year or not. If you really want to hold us to account, you can go print out Movelor’s current ratings and compare them for these teams at the end of the year.
Overperformer: Tennessee
Record against Movelor’s spread: 4–0
Opponents: Chattanooga, NC State, Kent State, Oklahoma
Net Movelor change since preseason: +8.4 points
Current Movelor ranking: 6th
Is Movelor still underrating Tennessee? It’s possible. Tennessee has sure played well so far, and I’m not sure their game against Oklahoma was as close as the final score. Still, I think Movelor has probably caught up to the Volunteers, or will at least catch up soon.
Two reasons for that:
First, NC State is a big underperformer of almost everybody’s expectations. Between them, Oklahoma, and Kent State—who is the worst team in FBS football—Tennessee’s drawn some opponents who were probably overvalued by Movelor and other systems entering the year.
Second, we have those concerns we talked about yesterday regarding Tennessee’s depth. We’re a long, long ways away from incorporating a team’s health into Movelor’s ratings. It does matter, though, and injuries tend to accumulate as a season goes on. You can get into a Ship of Theseus discussion pretty quickly here, so we’ll try to avoid going too far down that path, but we think Tennessee will peak earlier than Texas, Alabama, or Georgia, and that those three will maintain their current gap over the Volunteers.
Guess: Properly rated by Movelor.
**
Enjoying what you’re reading? Subscribe to our Substack to receive it every day in your email inbox.
**
Underperformer: Michigan
Record against Movelor’s spread: 0–4
Opponents: Fresno State, Texas, Arkansas State, USC
Net Movelor change since preseason: –9.0 points
Current Movelor ranking: 10th
We knew Movelor was overvaluing Michigan entering the season. That was a big driver of our enhanced offseason process this year. Given that, and given Michigan’s still ranked tenth in the country despite not looking remotely capable of competing, at home, against a top-four team? We expect Michigan to fall further. We don’t think the bottom will fully fall out—Alex Orji and the Alex Orji playbook appear to be a stabilizing force—but it’s doubtful Michigan’s really a top-ten outfit.
Guess: Still overrated by Movelor.
Overperformer: Miami (FL)
Record against Movelor’s spread: 4–0
Opponents: Florida, Florida A&M, Ball State, USF
Net Movelor chance since preseason: +9.4 points
Current Movelor ranking: 12th
Miami. They’re hot! Cam Ward’s the Heisman favorite! A lot of people have them winning the ACC!
Miami might win the ACC, and Cam Ward might win the Heisman, and Miami might be a top-ten team. Why, though, was this the year Mario Cristobal finally figured it out? There might be a great answer to that, and I’m not saying that sarcastically. There really might be a great answer. But every one of Cristobal’s Miami rosters has ranked between 12th and 14th on the 247Sports Talent Composite, and they’re 16–13 overall in his 2.3 years. I just don’t believe they’ve transformed to the extent of actually being one of the ten best teams in the country.
Guess: Properly rated by Movelor.
Overperformer: USC
Record against Movelor’s spread: 3–0
Opponents: LSU, Utah State, Michigan
Net Movelor change since preseason: +6.5 points
Current Movelor ranking: 13th
I don’t think Movelor’s overrating USC. USC is good. This is very clearly the best USC team so far under Lincoln Riley, because (as we discuss ad nauseam) the 2022 USC team was much worse than its record. But like Miami, it’s hard to believe they’re over the hump just yet. Combine that with the teams they’ve beaten, who are a combined 2–10 against Movelor’s spread, and it seems like some of USC’s 6.5-point improvement could have probably been assigned to Michigan, LSU, or Utah State.
Guess: Properly rated by Movelor.
Overperformer: Louisville
Record against Movelor’s spread: 3–0
Opponents: Austin Peay, Jacksonville State, Georgia Tech
Net Movelor change since preseason: +6.3 points
Current Movelor ranking: 20th
There are two kinds of ACC and Big 12 teams: Those with talent, and those without it. Clemson, Florida State, and Miami are in the Talent Composite top twenty. Nobody else is, and no one from the Big 12 is in the top 25. Still, programs like Utah (33rd in the talent composite), Louisville (36th), and Kansas State (64th) are clearly top-25 teams. What am I saying? I’m saying Louisville’s kind of like the Big 12’s powers. Jeff Brohm is getting a lot of quality for his talent.
I think Movelor’s still sleeping on the Cards. It’s probably overvaluing Notre Dame (whom it has ranked 8th), but there’s still at least a one-touchdown gap between where Movelor has the ND/Louisville game and where betting markets have it. Louisville really might be a top-15 team.
Guess: Still underrated by Movelor.
Underperformer: Florida State
Record against Movelor’s spread: 0–4
Opponents: Georgia Tech, Boston College, Memphis, Cal
Net Movelor change since preseason: –9.4 points
Current Movelor ranking: 24th
I wish Florida State wasn’t in Movelor’s top 25. I can explain why they’re there, but it’s a hard explanation. With so few people knowing what Movelor is, Florida State keeps us from listing Movelor’s top 25 on social media. Maybe we’ll make a top-23 graphic for Thursday. We cannot show a hostile internet whom we have rated as the 24th-best team in the country.
What’s happened with FSU? They’re underachieving what they should be. It’s not just what Movelor thought they would be. It’s what they should have been.
I’m extremely curious what happens when FSU plays SMU on Saturday in Dallas. My instinct says FSU will win and will right this ship and will not actually be worse than the 36th-best team in the country. (SMU’s rating has only changed by 0.4 points from what it was preseason.) It stinks to use instinct, though. The point of building a model is to avoid trying to grade your instinct.
Guess: Properly rated by Movelor but I’m going to barf now.
Overperformer: UCF
Record against Movelor’s spread: 3–0
Opponents: New Hampshire, Sam Houston, TCU
Net Movelor change since preseason: +5.1 points
Current Movelor ranking: 25th
Sam Houston’s rating is up 5.4 points since we started the season, the Bearkats settling in after a rough entry to the FBS world last year. UCF still covered against them. I was extremely skeptical of the UCF hype entering the year, and the fact they needed such a big comeback against TCU isn’t a great sign, but I think these guys still have more ground to gain.
Guess: Still underrated by Movelor.
Underperformer: Arizona
Record against Movelor’s spread: 0–3
Opponents: New Mexico, Northern Arizona, Kansas State
Net Movelor change since preseason: –6.6 points
Current Movelor ranking: 29th
Like Michigan but to a lesser extent, we had Arizona in mind when we updated our offseason ratings-changing process. There was a lot of roster continuity, but the head coaching change was transformative.
Arizona is not bad. But this team has shown us serious issues in all three of its games, and it’s shown different serious issues each time. It’s not just that they haven’t met the expectations. It’s the manner in which they’ve missed them, methods which have ranged from terrible tackling to offensive impotence to an all-around underwhelming effort.
Guess: Still overrated by Movelor.
Overperformer: BYU
Record against Movelor’s spread: 4–0
Opponents: Southern Illinois, SMU, Wyoming, Kansas State
Net Movelor change since preseason: +9.3 points
Current Movelor ranking: 30th
Our first hint BYU was good was when they walloped SIU, who’s nationally competitive at the FCS level. Saturday night, though, was the real breakout. BYU is in the spotlight, and they’ve earned it.
That said…
SP+ has a feature where it lists an “adjusted margin” for every game. This comes from a collection of stats which are very predictive of future final scores. More predictive than what we use! SP+ is more accurate than Movelor, at least at this point in their respective histories. Those stats say BYU should have only won Saturday night’s game against K-State by 11.3 points, not 29. Objectively, this team probably isn’t going to keep covering spreads forever.
Guess: Properly rated by Movelor.
Overperformer: Rutgers
Record against Movelor’s spread: 3–0
Opponents: Howard, Akron, Virginia Tech
Net Movelor change since preseason: +3.7 points
Current Movelor ranking: 34th
Rutgers? The 34th-best team in the country?? Believe it. We’re even asking if they might be better than that.
For all the Greg Schiano Piscataway-specific magic, though, we don’t think Movelor’s very wrong about Rutgers. It’s missed their final scores by 1.1, 6.5, and 6.0 points. Rutgers is on this list through a statistical oddity. Movelor seems to have a great read on how good they really are.
Guess: Properly rated by Movelor.
Overperformer: Cal
Record against Movelor’s spread: 4–0
Opponents: UC Davis, Auburn, San Diego State, Florida State
Net Movelor change since preseason: +5.1 points
Current Movelor ranking: 38th
Again with this one, the change in rating isn’t very big on a per-game basis. Again, Cal played a team Movelor at least initially overrated. Credit to Cal, but we think Movelor has found their level. It’s a good level! But it shouldn’t change much from here.
Guess: Properly rated by Movelor.
Underperformer: Kansas
Record against Movelor’s spread: 0–4
Opponents: Lindenwood, Illinois, UNLV, West Virginia
Net Movelor change since preseason: –5.4 points
Current Movelor ranking: 47th
Sometimes when a team blows a lot of leads or comes back a lot, statistical thinkers undervalue the comeback. They say, “For 80% of the game, Kansas was a lot better than West Virginia,” but they neglect how much better West Virginia was for the final 20%. With KU, though, I do think there’s been some flukiness to how they’ve underperformed. And like Rutgers and Cal, Movelor hasn’t actually been that wrong about them on a per-game basis.
Guess: Properly rated by Movelor.
Overperformer: UNLV
Record against Movelor’s spread: 3–0
Opponents: Houston, Utah Tech, Kansas
Net Movelor change since preseason: +6.8 points
Current Movelor ranking: 51st
Going hand in hand with that…I like UNLV a lot and I also don’t think they’re that good. Some of this is funky stuff that happened against Kansas. Some of it is how bad Houston’s offense is. Some of it is SP+ and FPI both doubting the Rebels more than Movelor does.
Guess: Properly rated by Movelor.
Underperformer: Wisconsin
Record against Movelor’s spread: 0–3
Opponents: Western Michigan, South Dakota, Alabama
Net Movelor change since preseason: –4.7 points
Current Movelor ranking: 59th
I don’t know that we yet know very much yet about Wisconsin. They played a MAC school, a good FCS team, and then Alabama in a game where the Badgers lost their starting quarterback early. I don’t think they’re in a good place—there were tons of concerns about the offense even before Tyler Van Dyke went down—but I’m not sure they’re any worse than what we’ve seen so far. It’s Wisconsin, right? They shouldn’t be that bad.
Guess: Properly rated by Movelor.
Underperformer: UCLA
Record against Movelor’s spread: 0–3
Opponents: Hawaii, Indiana, LSU
Net Movelor change since preseason: –5.5 points
Current Movelor ranking: 61st
UCLA came very close to missing this list. LSU covered Movelor’s spread on Saturday by 0.3 points. This is not a team where Movelor’s whiffing every Saturday.
Guess: Properly rated by Movelor.
Underperformer: NC State
Record against Movelor’s spread: 0–4
Opponents: Western Carolina, Tennessee, Louisiana Tech, Clemson
Net Movelor change since preseason: –8.1 points
Current Movelor ranking: 63rd
To Movelor’s credit, Movelor is not the one who had NC State in the top 25 to enter the year. It too, though, has been surprised by just how bad this team has been.
We think that’s meaningful. Preseason expectations come from somewhere, and the preseason AP poll is arguably the only AP poll governed by any sort of logic.
Guess: Properly rated by Movelor.
Overperformer: Arizona State
Record against Movelor’s spread: 4–0
Opponents: Wyoming, Mississippi State, Texas State, Texas Tech
Net Movelor change since preseason: +7.0 points
Current Movelor ranking: 69th
This is a tough one. On one hand, Movelor definitely underestimated ASU. On the other, it was only 1.4 points off the final score on Saturday in Lubbock. Texas Tech has been very tough to predict so far, missing in a lot of different directions, so I’m not sure how much we should make of that.
What I do think we should look at is where Movelor now has ASU within the Big 12: 14th, ahead of only Cincinnati and Houston. This is getting back to instinct, but…that doesn’t sound very far from the truth, especially since they’re within two points of Baylor and Colorado. They’ve made progress, but I think we’ve learned who they are.
Guess: Properly rated by Movelor.
Underperformer: Northwestern
Record against Movelor’s spread: 0–4
Opponents: Miami (OH), Duke, Eastern Illinois, Washington
Net Movelor change since preseason: –6.1 points
Current Movelor ranking: 71st
Northwestern has scored 69 points this season. They’ve scored 38 against teams in the FBS. (EIU is not a good FCS team.) Of those 38, seven came in overtime against Duke and two were scored on a safety against Washington. Through three FBS games, then, Northwestern’s offense is averaging 9.7 points per game.
That’s a weird number. Even for a bad offense, that’s a weird number. We don’t expect Movelor to be wrong now on Northwestern, but I’m not expecting the Wildcats to be rated much worse than this when all the scores have been counted.
Guess: Properly rated by Movelor.
Overperformer: Stanford
Record against Movelor’s spread: 3–0
Opponents: TCU, Cal Poly, Syracuse
Net Movelor change since preseason: +4.8 points
Current Movelor ranking: 97th
Movelor hasn’t been missing by much on Stanford on a per-game basis. I do not, though, think it’s done missing yet. It currently has the Cardinal only half a point better than the worst team in the Power Four. That seems like overkill. They belonged on the field with TCU and Syracuse, and while TCU and Syracuse aren’t world-beaters, that’s still not nothing. Stanford will not cover in perpetuity, but they should finish the year very comfortably within Movelor’s top 100. (There are five FCS teams ahead of them, for those wondering.)
Guess: Still underrated by Movelor.
Underperformer: Wake Forest
Record against Movelor’s spread: 0–3
Opponents: North Carolina A&T, Virginia, Mississippi
Net Movelor change since preseason: –4.1 points
Current Movelor ranking: 99th
Is Wake the worst Power Four team? Movelor says yes, but I think it’s Purdue (who covered against Indiana State, thereby avoiding the list). Between that, the small sample, and the small per-game miss, I don’t think Wake should worry that Movelor’s still too high. They found their bottom.
Guess: Properly rated by Movelor.
**
Movelor’s Week 5 picks are up, in addition to our bracketology, playoff probabilities, and Week 4 recap. We’re getting into the portion of the season where the model’s pretty accurate, so if you want to form opinions based on what the model’s saying, we won’t discourage you beyond what we said about Michigan, Louisville, UCF, Arizona, and Stanford (and a few other teams, I suppose). The only thing that has us holding our breath from here is the initial CFP rankings, something which tends to push some teams down and others up in a systemic, this–error–shouldn’t–exist–in–our–model way. But we don’t have to talk about those for weeks. We’re riding high, friends.
**
Thanks for being here. If you like what you just read and want it delivered to your email inbox every morning, subscribe to our Substack.
