How the Zeitgeist List works.

What the list measures

The Zeitgeist List ranks all 138 FBS programs by stature: the overall standing of the program, career-arc, weighted toward the present. It is not a power ranking (“how good right now” is our power model’s job) and not a history award (“how good you once were” lives in the Heritage badge). It answers a different question: who are the sport’s real institutions today, given everything they’ve done — heavily favoring recent results, but genuinely honoring the full arc.

Two design commitments shape everything below. First, transparency: every input is published, every constant is frozen and versioned, and a fan with a napkin should be able to check the math. Second, actual success has teeth: winning things counts co-equally with being good, so an erratic three-time champion can rank above a steadier program with no rings.

The formula

Three weighted pillars and two small channels that live outside the weights. Performance and Achievement are deliberately co-equal: how good you’ve been, and what you actually won, each carry 45%.

Performance — 45%

How good the program has actually been on the field, from our opponent-adjusted power rating (methodology here). Each completed season contributes its final rating, scored against that season’s FBS field — so a great 2019 team is measured against 2019, not against today. Seasons then fade on a three-year half-life: last season counts fully, a season three years back counts half, six years back a quarter.

An in-progress season enters in proportion to games played — a September résumé carries September weight, and in the offseason the not-yet-played season carries essentially none. The power model also rates FCS and lower-division teams on the same scale (anchored by the cross-division games that actually get played), which matters for programs that moved up — more on that below.

Achievement — 45%

What the program has won. Every trophy-shaped event earns points on a fixed schedule:

Why playoff depth, not appearances. The playoff grew from four teams to 12, so counting berths would inflate the modern era. Scoring by furthest round reached is era-proof: only four teams reach the final four in either era, and a four-team-era berth was the semifinal, so it automatically earns the final-four points. A 2015 playoff berth is worth more than a 2026 first-round exit — on purpose.

Why conference titles are multiplied. A title is priced by how hard the league actually was that season, measured from the league’s own power ratings (excluding the champion, so a great champion doesn’t inflate its own trophy). In 2024 an SEC title was worth about 4.3 points and a Mountain West title about 2.2. The multiplier is pinned to the season it was earned — a title won in a conference that later collapsed keeps its original value.

How history fades. Recent achievements decay on a five-year half-life; deep history automatically shifts to a slower, roughly 13-year half-life. The result: a 2024 national title is worth about 11 times a 1995 one, but the 1995 one never hits zero. And achievement is scored on an absolute scale against a fixed anchor — a title is worth a title regardless of who else is winning — with the anchor set so that even a maximal trophy case can’t quite peg 100. That anchor is what stops decades of ancient hardware from re-crowning a program that stopped winning.

Actually beating someone matters

Opponent-adjusted ratings have a known blind spot: they barely distinguish a one-point win from an overtime loss. The margin was the same; the outcome wasn’t. Two small channels outside the weighted core add the discreteness back:

• Signature wins — a bonus for beating genuinely strong opponents, priced by how strong the opponent was the week the game was played (locked forever — no re-pricing a September win because the opponent collapsed in November). Capped at +3 total, and wins that already minted a trophy above — a playoff win, a conference-title-game win — are excluded so nothing is counted twice.
• Bad losses — a narrow debit for losing to clearly below-median opponents, capped near −1. It’s deliberately hard to trigger: most famous upsets are losses to average teams, which the performance pillar already prices. Losses to lower-division opponents in guarantee games are exempt — that result is already a ratings catastrophe, and the formula doesn’t editorialize on top of it.

Talent — 10%

A blend of recruiting (60%, the industry composite) and NFL draft capital produced (40%). It’s deliberately the smallest weight: talent is a leading indicator of what a program might become, and a stature list is mostly about what a program is and has been.

Slow tier, fast outlook

The score recomputes weekly, but a program’s tier is deliberately sticky — a credible institution ranking shouldn’t whipsaw on one Saturday. Two mechanisms slow it down:

• A ±0.5 buffer on every tier line. Crossing a line puts you in range; you have to clear it by half a point before the tier actually changes. In plain terms: the list has to be sure before it moves you.
• A one-week confirmation on win/loss swings. The signature-win and bad-loss channels can move a score more than the buffer in a single game, so a tier change driven by one result must hold through the next weekly recompute before it lands. The score moves immediately; the tier waits a week.

An honest consequence: the tier is path-dependent. Two programs at the same composite score can legitimately sit in different tiers depending on where each came from. The score→tier map is not a clean lookup table, and we’d rather say so than pretend otherwise.

The fast layer is the Outlook: each week every program is marked climbing, holding, or sliding, with its literal distance to the next line (“+1.8 above the Power Program line”) and any gate countdowns. That’s where the week-to-week story lives, so the tier doesn’t have to tell it.

The title gates

The top two tiers require hardware, not just a score:

• Dynasty — two national titles in the trailing three seasons, or three in the trailing seven.
• Blue Blood — at least one national title in the trailing 12 seasons.

Titles only — a championship-game appearance does not count. A program can post the No. 1 composite score in the country and still be held at the tier below because it hasn’t won the thing; that’s the gate working, not a bug. It also means the Dynasty tier can be legitimately empty when no true dynasty is active. Gates expire on a calendar: when a title is one season from aging out of a program’s window, the list says so in advance (“Dynasty status expires after 2029 unless they win again”) — a demotion driven by the calendar is never silent.

The ladder

Fourteen tiers in five named bands: Dynasty (tier 1), Blue Blood (2), Power Program I–III (3–5), National Program I–III (6–8), and Regional Program I–VI (9–14). The cutlines are fixed values, not percentiles — calibrated once against the real score distribution, then frozen — so “Tier 6” means the same thing in 2019 and 2030, and a whole era can genuinely rise or fall together.

Band names describe caliber, not this season’s situation — and they are computed merit, not conference membership. A strong Group of Five program can reach Power Program by earning it. Separately, a Heritage badge marks what a program has been historically; Nebraska carrying a blue-blood-legacy badge at a modest live tier is the distinction working as intended.

Programs that came up from FCS

A program like North Dakota State arrives in FBS with a dynasty’s worth of FCS history and almost no FBS résumé. The list doesn’t pretend that history never happened, and it doesn’t overpay for it either:

• Performance carries over automatically. The power model rates every division on one scale, bridged by the cross-division games that actually get played, so FCS-era seasons already flow into the performance pillar at their honestly-discounted level.
• An FCS national title is priced by measured difficulty. It earns conference-title-style points multiplied by the measured strength of the 24-team FCS playoff field. In 2024 that priced an FCS title at about 2.2 points — between an American and a Mountain West title — computed from the bridge games, not asserted.
• Short-history shrinkage counts rated seasons, not FBS seasons. A program with 12 well-rated FCS seasons is not treated as a one-year-old data point; a genuinely thin two-season program still gets pulled firmly toward the middle.

The intended landing spot — and where NDSU actually lands — is “solid national program”: off the bottom, nowhere near the top.

What’s still settling

The formula constants above are frozen and versioned, and every recalibration ships as a published methodology event — never a silent change. The current tier floors are published but not yet formally frozen; a recalibration shuffles neighbors, not strata. The weekly in-season path is live, so the board updates through the season, and every weekly result for every program is kept permanently in a ledger with its raw inputs — which is also why past seasons on the board are restated under the current methodology, so years compare cleanly instead of preserving old formula quirks.