College Football Power Ratings: The Market’s Number for Every FBS Team

A ranking tells you someone thinks Georgia is the third best team in college football. A rating tells you Georgia is worth a certain number of points against an average team, which is the only version of that opinion you can actually bet. That distinction is the whole page. For Week 3, the board is fitted to 157 posted lines, and Indiana sits at the top with a rating of +33.4, which means something specific once you know how to subtract it from another team’s number.

Every rating here comes from the market itself, not from a poll or an editor’s gut. Home field this season is measured at 2.4, and that figure gets folded into every projected spread on the slate below. Subtract two ratings, add home field, and you have the same math a sportsbook used to set its line. Read the table with that in mind and it stops being a list of teams and starts being a tool.

College football Power Ratings: Week 3

A Number in Points Beats a Number on a List

Every college football media outlet runs some version of a power rankings column, and every one of them is unbetable by design. An editor moves a team from 11 to 8 because it won on the road, and that tells a reader nothing about how many points separate two teams or which side of a number holds value. It is an opinion dressed up as order.

A rating fixes that by putting a number on it. Indiana at +33.4 is not just ahead of the pack, it is a specific distance ahead of the pack, and that distance is what turns into a spread once you bring another team’s number into it. A ranking answers who is better. A rating answers by how much, and by how much is the only question a spread cares about.

That matters more in college football than almost anywhere else, because the gap between the top of FBS and the bottom is enormous. Indiana and Massachusetts are separated by 65.5 points on this board, a spread no ranking could ever express. A list just says one is better. The rating says how much better, which is the number you actually need before you touch a game.

Once you accept that a rating is a tool and a ranking is a take, the rest of this page is just learning to use the tool. That starts with the simple arithmetic behind every line in college football.

The Arithmetic Behind Every Spread

The formula behind this board is not complicated. Take the home team’s rating, subtract the away team’s rating, add home field, and that is the spread the market’s own numbers would produce. Home field this season is fitted at 2.4, and that figure comes out of the same lines as everything else on the board, not from a textbook assumption.

In college football that home field number carries less weight than it might seem, because the ratings themselves run so wide. A 30-point favorite is routine in this sport in a way it never is in the NFL, so a couple points of home field rarely flips a game the way it can in a tighter league. It still matters on the margins, especially in conference games between teams that are actually close.

What matters more than a half point of home field is where a number sits relative to a key number. A line that moves from 2.5 to 3, or from 6.5 to 7, means more to how a game gets bet and graded than a swing of a full point somewhere out at 24. Scoring is spread wider in college football than in the NFL, so those key numbers do not dominate the way 3 and 7 do on Sundays, but they still cluster more finishes than a random distribution would.

So when you use this table to build your own number, do the subtraction, add 2.4, and then ask where that number lands relative to the nearest key number before you decide whether the gap against the posted line is worth caring about.

Where These Numbers Actually Come From

These ratings are not a projection built from stats or a formula I invented. They are fitted, using least squares, to every point spread the sportsbooks have posted for FBS games this season, currently 157 lines. The fit finds the one rating per team, plus one home field number, that best explains every line the market has already set.

That makes this a market-implied rating. It is not my opinion about who is good, it is the market’s opinion about who is good, written down as a single number per team instead of scattered across a season’s worth of individual lines. When I say a team is rated at a certain number, I am saying that is what the combined judgment of every book and every bettor moving those lines currently believes.

Home field works the same way. Instead of assuming a flat number because that is what everyone has always used, the fit measures it from the same data, currently landing at 2.4. Neutral-site games carry no home field term at all, since there is no home team to credit. That measured approach matters more in college football than in pro leagues, because the gap between a true home game and a neutral-site bowl or a Week 0 kickoff game is not something you want to guess at.

The result is a board that reflects what the market has actually priced, not what a computer thinks should happen. That is also why the ratings move only when the market moves, which is the subject of the next section.

What a Weekly Move Actually Tells You

A media power ranking moves because a team won or lost last Saturday. These ratings move because the market changed its mind about a team, and those are not the same trigger. A team can win by three scores and barely budge here if the market already expected exactly that. A team can lose and hold steady if the market already knew it was overmatched.

This week’s clearest example is BYU, which moved +17.1 points up in the fit. That move is not a reaction to a highlight, it is the market repricing that team across every line it touches, all at once, because new information (An injury, a personnel change, a string of results) got absorbed into the numbers.

The reason this matters for betting is timing. The market moves before the scoreboard does. By the time a team’s record reflects a real change in quality, the lines already moved weeks earlier. Watching the weekly move column is watching the market’s opinion update in real time, which is a much earlier signal than waiting for a win-loss record to catch up.

That is why the movers are the first thing worth reading on this page every week, before the ratings themselves. A big move tells you where the market’s attention went. What it does not tell you is whether the market was right to move, which is the next question.

Checking the Market’s Homework

Next to the market rating sits a second number, the results rating. It is fitted the same way, using least squares, but instead of being fit to point spreads it is fit to what actually happened on the field, with margins capped at 24 points so one lopsided score does not distort a team’s whole season.

The gap between those two numbers, market minus results, tells you which teams the market is carrying above what they have actually produced, and which ones it is still underrating despite their results. This week that gap runs widest on Kentucky, rated 27.1 points above what its results support, and Massachusetts, rated 46.2 points below what its results support.

In general, trust the market number over the results number. The market has priced in schedule strength, injuries, and context that a raw margin-based fit cannot see, and it has done so with real money attached. But the exceptions are worth understanding rather than dismissing. A team can be unlucky in close games, can have played a schedule that inflates or deflates its results, or can have gotten a quarterback back that the results have not caught up to yet.

Read this column as a checklist of teams to look closer at, not as a signal to fade the market outright. When a gap is large and you can point to a specific reason for it, that reason is worth more than the gap itself.

Reading the Slate Without Overreading It

Every week’s slate lines up the spread these ratings produce for each game against the spread the market actually posted. The difference between them is a gap, and a gap means the market moved off its own season-long number for that specific matchup. That happens for real reasons: an injury, a weather forecast, a look-ahead spot, a public team drawing extra money.

Before Week 4, treat every gap as soft. A fit built on one or two games per team is a fit built on almost nothing, and a rating that thin can be knocked around by a single fluky score. The gaps do not go live in any meaningful sense until there is enough of a season behind each team to trust the number the fit is working from.

This week’s widest gap sits in North Dakota State at Sacramento State, where the posted line differs from what the ratings would produce by 3.2 points. That gap is worth a look, not a wager on its own. It says the market and the historical rating disagree about that specific game, and your job is to figure out which one has the better information, not to bet the gap blindly.

A gap is not a pick and it is not an edge by itself. It is a flag that says something changed between the season-long number and this week’s line. Sometimes that something is real and tradeable. Sometimes it is just a public team getting bet regardless of the number. The gap tells you where to look, not what to do.

My Tuesday Routine With This Table

I read the movers first. Before anything else, I want to know what the market learned in the last week, because that tells me where attention and information are concentrated. A team jumping several points in either direction is a team where something changed, and I want to know what before I look at anything else on the board.

Next I go to the market-minus-results column. That tells me which teams the market is carrying on reputation or schedule and which ones it is still sleeping on relative to what they have actually done. I am not betting off that column directly, but it tells me where to spend my time doing actual homework.

Then I go to the slate and look at the gaps, the games where the posted line has drifted furthest from what the season-long ratings would produce. I build my own number the same way the model does, rating minus rating plus home field, and compare it to the posted line. I only care about the difference once it crosses a key number.

None of this replaces watching tape or knowing injury news. A quarterback change breaks this model until the market has time to reprice it, and early-season numbers are soft by design because there is not enough data yet. A rating describes what the market currently believes. It is not a prediction of what will happen on Saturday, and I use it accordingly, as a starting point for my own number rather than a finished answer.

Get Jimmy's picks for free!

You'll get free picks via email, plus access to exclusive discounts!

<<

Sign Up

We won't spam you - ever. No obligation. You can unsubscribe at any time.