SPORTS.PICKS1024
Scores

No dated games available in this filter.

EXPLAINER · COLLEGE · NOT A PICK

How McIllece Sports Rates College Football Teams and Players in the Portal Era

SPORTS.PICKS1024 · with notes from Justin at McIllece Sports · Oct 6, 2026

The ratings, systems and College Football Atlas described here belong to McIllece Sports (mcillecesports.com). Facts come from the McIllece Sports website as of October 5, 2026, plus notes from Justin at McIllece Sports.

Every offseason now starts with the same question: thousands of players changed teams, so what is anybody's roster actually worth? The old preseason shortcut was counting returning starters or the share of last year's yards coming back. McIllece Sports decided in 2022 that the shortcut had stopped working and rebuilt its prediction methods around the transfer portal and NIL.

Here's a plain-English look at how its player ratings, power ratings and preseason magazine fit together.

Player ratings that travel with the player

The core stat is NetWin. Per the site, it's a player's estimated net win value as a full-time starter on an average FBS team running standard college football schemes, scaled to a 12-game season. Below-average players can have a negative NetWin.

That "standard schemes" baseline matters. The site's example is Navy quarterback Blake Horvath, who it says is worth much more in the Midshipmen's option system than he would be in a typical FBS offense. Rating everyone against the same baseline lets you compare players without tying them to one coaching staff.

Then comes the portal twist, and it's worth being clear about where it happens. The Portal Ratings use that same standardized NetWin, not team-specific numbers. Justin at McIllece Sports says that's on purpose: coaching staffs can change a fair amount between the portal window and fall camp, and he wanted the portal ratings to be a purely fair comparison of talent lost and added. The system adjustment comes later, in the Atlas preseason predictions, where he says it's crucial. That's where the site's "flexible, portable player ratings that are matched to the coaching systems of the destination programs" come in.

Here's how the matching works. He builds offensive, defensive and special teams systems for each program, weighted by how much each coach is responsible for. The head coach affects all three, the offensive coordinator affects every offensive position, and the running backs coach only affects the running backs. To measure how much each position matters to a staff (split into rushing and passing on both sides of the ball), he compares the staff's weighted history of win-probability effects against unit efficiency. Efficiency here is the sum of positive EPA divided by the sum of absolute EPA, using his in-house expected points added model. That shows which units, like quarterback pass offense or linebacker run defense, matter most to whether a staff wins or loses games. Put that together with his ratings for every player with experience in those categories, plus priors for program talent and coaching effects, and he can judge how a transfer fits his new scheme.

The public tables add some context. Pctile ranks a player against his position on a 0 to 100 scale, and players with little experience get pulled toward the middle because there's not much data on them. CGE (Complete Game Equivalents) measures experience as the total win-probability impact a player has had over his career. The site is blunt that "no single statistic tells the whole story."

Those standardized numbers roll up into the Portal Ratings, which rank all 138 FBS teams by the NetWin value of transfers added minus transfers lost. Only positive values count, so there's "no addition by subtraction" and no penalty for adding a player. A separate Player ($) Valuation page turns the ratings into estimated fair-market values, which McIllece stresses are not predictions of what anyone is actually paid.

Power ratings and the Atlas

Once games start, the Power Ratings page takes over. The π-Rating is the probability of beating an average team at a neutral site. Rather than starting from final scores, it tracks play-by-play changes in win probability and weights the moments that matter most. McIllece calls it "a power rating based on game control, not points."

The Power Rating turns that into points: the expected margin against an average FBS team on a neutral field. The site says it can predict the margin between any two Division I teams, with 2.0 points added for home field, and FBS and FCS teams share the same scale.

A Change column compares each team to its preseason number from the Atlas. When I checked on October 5, 2026 (the page was labeled 2026 Week 6), Georgia was No. 1 in FBS at 39.5, up 11.1 from preseason. Kentucky sat at 17.6, a jump of 16.0. The site even explains how to update an Atlas preseason line with this column: subtract the Change value for the team and add the Change value for its opponent.

Early ratings swing a lot, and McIllece says so. Preseason projections carry more weight early and fade as games pile up. From Week 5 on, the system adds iterative strength-of-schedule adjustments and weights games unequally to catch outliers like a fluky blowout. These ratings are predictive. For a ranking based only on what teams have accomplished, the site points to its CARAT Ratings.

All of this feeds the 2026 College Football Atlas, McIllece's 623-page preseason magazine. The current version 2 includes weekly prediction pages and player and coach indexes, and it gives each FBS team four pages, with odds and lines, over/unders, game and season predictions, player and coach ratings, portal rankings, roster dollar valuations and scheme diagrams, backed by 100,000 full-season simulations that include playoff odds game by game. New this year, per the site: bias-correction modeling, a "Postseason Progression" tool showing the records teams need to stay in the conference and playoff hunt, and general managers listed with coaching staffs.

McIllece says the Atlas has been the most accurate preseason magazine of the portal era, citing consensus tracking at preseason.stassen.com. You can get it at mcillecesports.com, and the homepage has a free sample of the Indiana pages if you want a look first.

The power ratings page is free and updates as the season goes, so it's easy to check your own picks against it week to week.

Sources

Back to news · College football