Papers · Preprint · v3

A Six Birds' Eye View of Dark Energy: Closure, Route Mismatch, and Audits for Apparent Acceleration

Can a lumpy universe fake acceleration?

In plain words

Cosmology fits smooth models to data that summarize a lumpy universe. So a constant term that appears in a fit may be a real component, or it may be standing in for the lumps.

The paper asks when a coarse description can have a law of its own at all. The exact answer, checked in Lean, is that micro states sharing an observable must evolve to states sharing an observable. Comparing average then evolve with evolve then average is the wrong test on full states: the two can differ even when an exact coarse law exists. Measured on the observables, the same comparison detects every failure. When the criterion fails, no correction term on those observables is exact, and a fitted term such as Λ is an effective description of the history actually realized.

Two toys show the mechanism. In one, the mismatch is exactly the time step times the nonlinearity times the variance. In the other, every patch decelerates while the average accelerates.

Synthetic distance data from a generator with no cosmological constant, fitted with the standard model, gives about sixty percent dark energy. A heterogeneity term fits just as well, and neither fit is statistically adequate. On public supernova and BAO data the pipeline recovers the published supernova parameters. The data do not prefer a correction that decays toward the past, and they do not single out a constant.

Three bars on one scale of misfit chi squared, for 40 synthetic distances and 2 parameters. A statistically adequate fit would score about 38. Lambda CDM scores 379.6 and the rewrite model 378.6, two almost equal bars ten times longer. Note: synthetic data from a toy with no Lambda; no claim that dark energy is an artifact.
Two models fitted to the same synthetic distances from the paper's toy, which contains no Λ. They fit equally well, and equally badly: with 40 points and two parameters an adequate fit would score about 38, and both score about 380. A fitted term that matches the data this way is a description of one history, not a detection.

What it shows

  • A Lean checked criterion for when a coarse law exists, with mismatch on the observables as the exact obstruction.
  • A generator with no Lambda that produces an apparent Lambda under fitting, as a mechanism.
  • A provenance tracked pipeline on public survey data, whose large scale structure runs are software validation only.

What it does not claim

It does not claim the cosmological constant is an artifact, and it proposes no microphysical alternative. The synthetic results establish a mechanism, not a property of the real universe. The toys are Newtonian and shear free, so they say nothing about the size of backreaction in the real universe. On background data a constant closure correction and a fundamental Λ are indistinguishable, and the audits that could separate them have not been run with physical theory codes.

Cite

Tsiokos, I. (2026). A Six Birds' Eye View of Dark Energy: Closure, Route Mismatch, and Audits for Apparent Acceleration. Zenodo. https://doi.org/10.5281/zenodo.23368267