Follow each test while it happens. Every mission starts with what we expect and what we do not know. The log records each step, from us and from the observatories and labs whose data decides the test. At the end, we count what we learned.
A result is one step in a mission, not its score. What counts is what we know at the end that we did not know at the start, including what we learned about ourselves.
What we expect and why, what we do not know yet, and the assumptions we are making.
Every change, review and outside data release, with the reason. Entries are added, never edited.
What we learned about the question, the data and methods, our process, and ourselves, and what carries forward.
Counted from the log below.
Newest first. Ic² entries are our changes. Outside entries come from the teams whose data decides a test. Each one says whether it came before or after the deciding data, because a change after the data deserves a closer look.
Version 1.1 of the record was deposited on Zenodo. It lists DR2 Results IV, fixes the 3.1σ baseline and the choice of CMB data, and gives 2027 as the expected release. The claim, the measure and the outcome boundaries are word for word the same, and version 1 stays available.
Why it matters: The review found gaps that were ours to close before the data, and closing them in public shows exactly what changed.
Record →The first eleven tests received mission names: the program's color and one of the 88 constellations, drawn from real star positions. COSMIC-007 is named for Orion because its predicted axis points there. Names are kept for good, whatever the result.
Why it matters: A name gives followers something to follow from start to finish.
Standard model comparison panels were added to COSMIC-005 and COSMIC-007, showing what ΛCDM predicts beside what the framework predicts.
Why it matters: A follower cannot judge a prediction without knowing what the alternative predicts.
A literature check on JWST's early massive galaxies found that a factor-of-10 claim in our audit was not supported by its source. Most current work explains much of the excess with bursty star formation. A caveat was added to the site.
Why it matters: COSMIC-003 was written after JWST's first results, so it was never a test. What we say about it still has to be accurate.
Fermilab says the first results from the full five-year survey are expected in 2027. The site had said from 1 December 2026. The date was changed on every page.
Why it matters: The deciding date belongs to the observatory, not to us.
The pre-test review found that the record did not list DR2 Results IV, public six weeks before the deposit. It fixed the 3.1σ baseline and the choice of CMB data in advance. The prediction and its boundaries did not change. Version 1.1 of the record is recommended.
Why it matters: A review before the data may complete a record, but it must never move the line between outcomes.
A claim audit compared the framework's public claims against current literature and listed what to watch.
Why it matters: Claims age. The audit sets a review date for each one.
Preprint version 1.1 was published on Zenodo with the corrected predictions and media table.
Why it matters: The public record should match what we now expect, and the earlier version stays available beside it.
Record →A model of the sound field inside the sphere showed that particles do not gather at the pressure node. Predictions 1 to 4 were corrected before any experiment: the shell's radius depends on the medium, and the next shells are not harmonics.
Why it matters: The original predictions rested on an assumption the physics does not support. Correcting them before any data keeps the test honest.
CYM-003, the crop formation comparison, was retired before any test. It kept its place on the results page as a record.
Why it matters: It assumed Chladni figures are slices of a three-dimensional sound field. They map how a flat plate vibrates, so the comparison had no physical basis.
COSMIC-007 was deposited on Zenodo. The decision rule adds a 3σ threshold and the specific axis to what the website first published, which makes it harder to satisfy.
Why it matters: A timestamped record, with the rule tightened and nothing relaxed.
Record →COSMIC-005 was deposited on Zenodo with its measure fixed in advance: DESI's significance for evolving dark energy, from BAO plus CMB without supernovae, must rise above about 3.1σ, with w₀ above −1 and wₐ below 0.
Why it matters: The framework's earlier dark energy claim was written after DESI's first results. This is its first formal prediction made before the data.
Record →DESI released DR2 Results IV, Lyman-alpha measurements from the same three-year data. DESI said they agree with the standard model and that the hints of evolving dark energy may fade, or that a more complex model may be needed.
Why it matters: It does not decide the mission, which waits for the five-year analysis, but it makes a partial or falsified outcome more likely.
The Vera C. Rubin Observatory began its ten-year Legacy Survey of Space and Time. The first full data release is expected in mid-2028.
Why it matters: The survey that decides this mission started after the axis was published.
The predicted axis, (l, b) ≈ (210°, −20°), was on the website by this date, the earliest the site's public history records.
Why it matters: The direction is the hardest part of the prediction to meet by chance, so its date matters.
DESI completed its planned five-year survey. The cosmology analysis of that dataset had not been released.
Why it matters: The deciding data now exists but is not public, so a prediction made now is still a prediction.
Source →The microgravity acoustic experiment and its predictions were documented, with preprint version 1.0 on Zenodo. CYM-003, a crop formation comparison, was documented at the same time.
Why it matters: Predictions are written down before any chamber is built or flown.
Record →The COSMIC Framework reference paper was deposited on Zenodo.
Why it matters: It fixes the framework's public form, so later predictions can be dated against it.
Record →What each mission expects, does not know and assumes. New missions will write this on the day they are registered.
Midnight EridanusDESI's five-year analysis shows evolving dark energy more strongly than the three-year data did, in the same direction.
Next pre-test review once DESI announces a date
Cerulean PhoenixCMB polarization shows a non-random pattern at an angular scale derived from the framework's information-density equations.
State the scale; check the dates on the card; deposit
Midnight OrionGalaxy distributions beyond about 100 megaparsecs are not the same in every direction, and the preferred axis points toward (l, b) ≈ (210°, −20°).
Pre-test review before DR1
Emerald LupusIn microgravity, particles in a driven sphere gather in three-dimensional shells and clusters set by the chamber's resonances, at radii that depend on the medium.
Pre-register flight geometries before flight
Citrine MonocerosFour effects on how people encode and remember information, beginning with a drop of at least 15% in comprehension when text asks for more than two or three working-memory chunks at once.
Align card and schedule wording; numbers for 1.2 to 1.4
Aqua DelphinusThe performance gap between biological and non-biological intelligence is structured by task type, because biological minds carry survival overhead that AI systems do not.
State the predicted pattern in numbers
Crimson CentaurusCP-violating particle processes release heat beyond momentum-transfer predictions, in proportion to the information erased.
Size of the effect; which data decides
Crimson DracoThe quark mixing angles and CP phase minimize an information-theoretic cost function.
Deposit the function before the comparison
Magenta AquilaEntanglement entropy at the confinement boundary scales with the framework's information-density parameter.
State the predicted scaling in numbers
Magenta SextansAt the quark-hadron crossover, entanglement entropy drops in a way that thermal predictions do not produce.
Fix the analysis and threshold before looking
Citrine TelescopiumAutophagy and synaptic pruning release heat consistent with Landauer's limit, about 10⁻²¹ joules per bit, detectable above baseline.
State the detection threshold
This checks how each test is set up, not how it turns out. A gap here is the next thing to fix.
| Mission | Written before the data | Timestamped record | Numeric threshold | Reviewed | Decided by | Next step |
|---|---|---|---|---|---|---|
| Midnight EridanusCOSMIC-005 | Yes | Zenodo, 12 Sep 2026; v1.1 17 Sep 2026 | YesAbove about 3.1σ; w₀ > −1, wₐ < 0 | YesPre-test review, 16 Sep 2026 | DESI five-year results, expected 2027 | Next pre-test review once DESI announces a date |
| Cerulean PhoenixCOSMIC-008 | To check | Website only, May 2026 | Not yetNo numeric scale stated | Not yet | Simons Observatory; no release date announced | State the scale; check the dates on the card; deposit |
| Midnight OrionCOSMIC-007 | Yes | Zenodo, 12 Sep 2026; website by May 2026 | YesAbove 3σ, axis within stated uncertainty | Not yet | Rubin DR1, mid-2028 | Pre-test review before DR1 |
| Emerald LupusCYM-002 | Yes | Zenodo preprint v1.0 Feb 2026; v1.1 15 Sep 2026 | YesMode frequencies and shell radii by medium | PartlyModel check, 15 Sep 2026 | Parabolic flight, after CYM-001 | Pre-register flight geometries before flight |
| Citrine MonocerosCOG-AUG I | Yes | Site record, 31 Jan 2026 | Partly15% for effect 1.1 only | Not yet | Internal study, testing from Feb 2027 | Align card and schedule wording; numbers for 1.2 to 1.4 |
| Aqua DelphinusNBI-005 | Yes | Site record, 2 Mar 2026 | Not yetNo numeric pattern stated | Not yet | Performance testing, from Apr 2027 | State the predicted pattern in numbers |
| Crimson CentaurusCOSMIC-SD-001 | To check | Site record, May 2026 | Not yetProportional, no number | Not yet | LHCb comparison, from May 2027 | Size of the effect; which data decides |
| Crimson DracoCOSMIC-SD-002 | Data already public | Site record, May 2026 | Not yetCost function not derived | Not yet | Published angles, once the function is fixed | Deposit the function before the comparison |
| Magenta AquilaCOSMIC-SD-003 | Yes | Site record, May 2026 | Not yetScaling stated, no number | Not yet | Lattice QCD from Apr 2027; EIC in the 2030s | State the predicted scaling in numbers |
| Magenta SextansCOSMIC-SD-004 | Data already public | Site record, May 2026 | Not yetNo size stated | Not yet | Existing RHIC and ALICE data, analysis from Feb 2027 | Fix the analysis and threshold before looking |
| Citrine TelescopiumCOSMIC-013 | Yes | Site record, 24 Jan 2026 | PartlyHeat per bit stated; detection level not | Not yet | Internal testing, from Apr 2027 | State the detection threshold |
The first closed prediction. It was retired before any test, and it still taught us something in every group.
Carries forward: Check each premise before registering a test. Retire a prediction as soon as its premise is shown not to hold, and keep the record in public.
The Quest
A short update every two months, and a note whenever a test is decided.