Only results count. No polls, no preseason rankings, no recruiting, no eye test, no margin of victory.
Every game is one line on a team's ledger. A win is worth +1 to +2 and a loss costs -1 to -2, depending only on how good the opponent is, and "good" is measured by the opponent's own ledger, all the way down. A team's rank is the sum of its lines.
Everything is recomputed from scratch every week, so a September win over a team that turns out to be good gains value in November.
Extra games are earned. Reaching a conference title game or the playoff earns a line, playoff wins earn more, and in the playoff beating a team puts you above them, no matter what the ledger says.
Every rank comes with a receipt. Open any team on any week's page and the math for every game is there.
A college football ranking with no polls, no preseason opinions, and no "eye test." Just results.
Everybody starts at zero. No preseason rankings, no recruiting classes, no "they were good last year." Week 1 is a blank sheet for every team.
Every game you play adds one line to your ledger. Win, you gain points. Lose, you lose points. Your ranking is your total. That is the whole system.
How many points depends on how good your opponent is, and "good" means exactly one thing: how their own ledger reads in their other games. Not their poll ranking, not their reputation, not even their record. Their ledger points per game in the games they played against everyone but you. And their ledger depends on their opponents' ledgers in the same way, all the way down.
Beat a team whose other games average the best possible +2.0 per game: +2.0 (the max).
Beat a team whose other games average the worst possible -2.0: +1.0 (the min).
Lose to the best: -1.0 (the least painful loss). Lose to the worst: -2.0 (the most painful loss).
In between it is proportional. For the math people: win = 1.5 + (opponent's figure)/4, loss = -1.5 + (opponent's figure)/4, where the figure is their ledger points per game in their other games. An average opponent, figure 0, makes a win worth exactly 1.5.
Because everyone's figure depends on everyone else's, the whole table is solved at once: start every team at average, recompute every line, repeat until nothing moves. It settles in about a dozen rounds, and it always settles on the same answer.
The game against you is never counted directly when judging that opponent. Through longer loops in the schedule a result can still echo back faintly; that is unavoidable once everything depends on everything.
How a line gets its value. Ironwood beat Copper Valley, so Ironwood's line depends on how Copper Valley did against everyone else, which depends on how those teams did against everyone else, and so on. The chain eventually loops back on itself, which is why the whole table is solved at once.
The whole thing is thrown out and recomputed every week. Nothing is locked in. A week 1 win over a team that turns out to be great gets more valuable in November. A week 1 win over a team that collapses gets less valuable. Nobody's opinion is ever consulted.
In week 1, everybody's opponent is average. Nobody has a ledger yet, so every opponent counts as a 0. Every winner gets +1.5, every loser gets -1.5, and half the country is tied for first. That is on purpose. Strength of schedule does not exist yet, so nobody gets credit for one.
FCS opponents count as one fixed opponent, the worst possible one. Beating an FCS team is worth exactly +1.0, the minimum a win can be worth. Losing to one costs exactly 2.0, the most a loss can cost. FCS teams have no ledger and are not ranked.
Extra games are earned, so reaching them counts.
Both teams in a conference championship game get a +1.0 line for reaching it.
Every team in the playoff field gets a +1.0 line for making it.
A playoff win adds a bonus on top: +0.5 first round (a bye counts, you earned it), +1.0 quarterfinal, +1.5 semifinal, +2.0 for winning the title.
Losing any of those games is a normal loss. The credit is for earning the game, so a team that loses a title game or a first-round playoff game no longer sits behind an equal team that stayed home, and because both teams get the line, a loser can never pass the team that beat them on the strength of that game. Regular bowls are normal games with no line.
A team's playoff lines appear once it has played a bracket game, so bye seeds get theirs after the quarterfinals.
In the playoff, beating a team puts you above them. Period. A team ranks above every team it beat in the bracket, and above every team those teams beat. This is not points; it is an ordering rule applied after the ledgers are built, the way head-to-head breaks a tie. A bracket has no loops, so it can never contradict itself, and it only asserts things that happened on the field. It guarantees the champion ranks above every other playoff team and the runner-up above everyone on its side of the bracket. Teams the bracket does not connect, including everyone outside the playoff, are ordered by ledger as usual. When this rule moves a team, both teams' ledgers say so and why.
Margin of victory does not count toward points. Winning by 50 is worth the same as winning by 1. Running up the score gets you nothing. Home, away, or neutral does not change the points either. Both only appear as tiebreakers, below.
If two teams have the exact same points, tiebreakers go in this order, and the first one that separates them wins:
Head-to-head (only if the results settle it cleanly; a three-way circle skips this)
Strength of schedule (average quality of everyone you played, the same number as rule 3)
Quality wins minus bad losses (wins over above-average teams, losses to below-average teams)
Average margin of victory, capped at 21 points per game
Record against common opponents
Road record (true away games only)
If they are still tied after all six, they share the rank.
Every ranking comes with a receipt. For every team you can see every game, the opponent's figure, and the exact math for that line, plus which tiebreaker or bracket rule placed them. If you think a team is ranked wrong, you can point at the exact line you disagree with.
Things this system never looks at: AP or Coaches polls, committee rankings, preseason anything, recruiting rankings, betting lines, last season, or who "looks like" a good team.
How this differs from the Colley Matrix
YARDS uses the same idea as Colley for judging opponents: every team's number depends on every other team's, solved together, from wins and losses only. It differs in what is counted and shown. YARDS is a sum, so a 13th game won is an accomplishment that adds to the ledger; Colley is a rate near 0.5 that a 13th game barely moves. YARDS never counts a game directly as evidence about itself; Colley counts your loss to a team when rating that team for the purpose of valuing that same loss. YARDS credits reaching and advancing in the playoff as stated rules; Colley has no notion of a tournament. YARDS shows every game's value in the same units as the rank; Colley's rating can be decomposed but nobody reads the pieces. And every YARDS dial is visible: a quarter of the opponent's figure, a fixed FCS value, a 21-point cap that only breaks ties, a named tiebreaker ladder.
These rules replaced the original set on 2026-09-25. The original rules and results are archived under YARDS v1, and the changelog lists every difference with the date and the reason.