44 tests · 30 agree · 0 disagree 14 indeterminate — shown, not hidden
Test 5 — The early universe, two ways — the timeline, computed twice: 44 tests
Two unrelated methods reconstruct the history of the universe: a "record-cost" reconstruction built on the framework's granularity axiom, and the standard hot-Big-Bang calculation. The two pipelines share no equations, no calibration, and no free parameters — and they land on the same timeline at every epoch that can be computed, from ~20 microseconds after the beginning to today's 13.8 billion years. Across 44 checkpoints: 30 agree, 0 disagree, and 14 are honestly indeterminate — quarantined in the open, with the whole denominator always shown.
What was done
The full history of the universe was recomputed from the framework's side — the cost of keeping a physical record as a function of cosmic time — and lined up, epoch by epoch, against the standard calculation. Each of the 44 checkpoints is published as its own row with both derivations shown side by side: what the record-cost route returns, what standard cosmology returns, what is measured, and an honest verdict from a fixed three-word vocabulary (agrees / indeterminate / disagrees). The headline row is the simplest one: both routes return today's age of the universe — 13.8 billion years — from completely different starting questions.
Open the full 44-row scoreboard — every derivation side by side
Two ways?
Yes — and genuinely two. The record-cost route starts from one question: when does a distinction become affordable enough to survive as a record? The standard route runs the established thermodynamics of an expanding hot plasma. Neither pipeline reads the other's equations, and neither is calibrated to the other's answers. Thirty independent agreements between two routes that share no free parameters is a consilience argument in its cleanest form: two surveyors who never spoke, drawing the same map at every point they can both reach.
One piece of history belongs on this page. An earlier version of this test reported a forward-cosmology pipeline that refused to run at all: it was mechanically gated to compute the dark-energy density natively before loading any observed value, and rather than peek, it reported itself blocked. That refusal was the right discipline, and it was resolved the honest way — not by relaxing the gate, but by settling what the demand actually was. By the framework's own unit-gauge doctrine, no naked dimensionful number is derivable from pure mathematics; at least one measured ruler is mandatory for any theory. So the dark-energy density Λ is charged openly as the fifth measured anchor (≈ (2.3 meV)⁴, honestly measured, never derived), and the two-route comparison runs with every anchor declared. The blind element survives exactly where it belongs: in the standing-bet rows below, where the record-cost route has staked predictions the measurements have not yet returned.
Honest result
| Checkpoint | Both routes return | Measured / bound | Verdict |
|---|---|---|---|
| Age of the universe | 13.8 billion years — from two different starting questions | 13.8 Gyr (Planck 2018) | Agrees |
| Effective neutrino species Neff | 3.044 — the famously precise, non-integer value, reproduced by the record-cost ledger | 2.99 ± 0.17 (Planck) | Agrees |
| First nuclei (~3 minutes) | The deuterium-bottleneck breakout at T ~ 0.08 MeV, t ~ 180 s — reached identically by both routes | Primordial D/H and ⁴He within 0.2–0.7σ | Agrees |
| Matter–radiation equality | zeq ≈ 3422 | Planck: 3402 ± 26 (~0.6%) | Agrees |
| CMB lensing amplitude | Parameter-free prediction of the lensing amplitude | Planck: 1.011 ± 0.023 — a >40σ-significance agreement | Agrees |
| Scalar tilt ns | ns ∈ [0.9643, 0.9679] | Planck: 0.9649 ± 0.0042 — bracketed | Agrees |
| Monopole relic abundance | Ωh² ≈ 2×10⁻⁶¹ — both routes independently crush the classic overproduction problem | No monopole ever detected | Agrees |
The 14 indeterminate rows decompose — and the decomposition strengthens the scoreboard rather than weakening it. Six are standing bets (tests 08, 09, 10, 12, 27, 37): rows where both routes already agree with each other and only the measurement is pending. The headline bet is the primordial tensor-to-scalar ratio, r ∈ [3.5, 36]×10⁻³ — comfortably under the current BICEP/Keck bound (r < 0.036) and squarely inside what the LiteBIRD satellite (~2030) can confirm or kill outright. The other eight are open for all of cosmology (tests 06, 17, 18, 24, 25, 26, 36, 40): baryogenesis, the identity of dark matter, lithium-7, S8, the Planck wall itself — questions no framework currently answers, kept on the board in the open rather than deleted from it.
At the wall itself the framework states its reading plainly: the Big Bang is the first moment any distinction could be paid for — the origin of recordable history, not an explosion inside a pre-existing time. "Before the Big Bang" dissolves the way "north of the North Pole" does, with no singularity required — and both dimensional analysis and record-cost bookkeeping stop at the identical Planck time. See the Early & Distant Universe overview for that account in full.
Data sources
- The full scoreboard of record: /early-universe/tests/ — all 44 rows, each with the record-cost derivation and the standard derivation shown side by side, the measured value, and the verdict.
- Measured comparisons — Planck 2018 cosmological parameters (age, Neff, zeq, ns, the lensing amplitude 1.011 ± 0.023); primordial-abundance measurements for D/H and ⁴He; the BICEP/Keck 2021 bound r < 0.036.
- The named future test — the LiteBIRD tensor-band measurement (~2030), which either confirms the standing bet r ∈ [3.5, 36]×10⁻³ or falsifies the framework's inflaton candidate on its own terms.
What this does not show
The 14 indeterminate rows are genuinely open — baryogenesis, dark-matter identity, lithium-7 (an ~3× over-prediction both routes share, because both use the same nuclear network there), S8 — and the scoreboard says so on its face rather than quietly shrinking its denominator. Several agreements share standard growth and nuclear machinery where the two roads necessarily coincide; the Neff = 3.044 row, for instance, reuses the same entropy-conservation identity as standard cosmology — the distinct claim is that the record-cost reading reproduces it from a completely different motivation. The six standing bets are bets: if the measurements come back outside the stated bands, the rows fail in public. And none of this is peer-reviewed. Showing the whole denominator — 44, not 30 — is the point: a cherry-picker shows you the numerator; this shows the exam.