USR-2026-0004 · Dark energy

Binned w(z) rises from -0.920 ± 0.040 (z = 0.0-0.4) through -1.078 ± 0.093 (z = 0.4-1.0) to -1.537 ± 0.282 (z = 1.0-2.5); the trend survives dropping all SNe but dies without BAO.

PROVISIONAL BAO-anchored trend — no independent σ quoted reproduced ≠ discovered re-run & confirmed 2026-07-09 (R0 — founder CI; independence pending)

Model-independent bins show w above -1 today and below -1 in the past. Removing every supernova leaves Δχ² = 8.6; removing BAO collapses it to Δχ² = 0.8 — BAO is the load-bearing dataset, stated openly.

Falsify-box — how to kill this claim

SURVIVED trend survives dropping ALL SNe

Verification record — every quoted number, re-run 2026-07-09

quantitymeasuredverdict
binned w, z = 0.0-0.4 -0.920 ± 0.040 MATCH
binned w, z = 0.4-1.0 -1.078 ± 0.093 MATCH
binned w, z = 1.0-2.5 -1.537 ± 0.282 MATCH
kill test: drop ALL SNe Δχ² = 8.6 — survives MATCH
kill test: drop BAO Δχ² = 0.8 — dies MATCH

Provenance

script: expansion_seam/run_prove_it.py
script status: PRESENT
datasets: DESI DR2 BAO · Pantheon+ · Planck CMB (compressed) · claim-scripts (downloadable from /data — run it yourself)
re-run: 2026-07-09

Edges

REPRODUCES → DESI DR2 evolving-DE picture in model-independent bins

Status, honestly

R0 pass 2026-07-09. Status here is computed from evidence — the author cannot set it, and neither can we. Independent reproduction would move it; nothing else will.

§ Figures 1 from expansion_seam/run_prove_it.py

What this claim's own script draws. These are the founder's committed outputs — the same plots the script regenerates on any machine that runs it.

prove_binned_w.png 63.7 KB download
§ Reproduction run_prove_it.py

The founder's own run of this script, captured verbatim. A match proves the result is reproducible; it is still R0 on this registry's independence rings — same code, so it cannot move a status. Only an outside run does that.

sha256 e24d96680083b7412d24cf71… size 8440 B exit 0 runtime 11.1s env python 3.12.2 · numpy 1.26.4
A · Recorded founder's machine · 2026-08-02
TEST 1 — w(z) in independent bins (no parametrization)
  best joint fit: w = [-0.92  -1.043 -1.336]   chi2=1413.7
  bin 1  z=0.0-0.4 :  w = -0.922 +/- 0.040   (2.0 sigma from -1, ABOVE)
  bin 2  z=0.4-1.0 :  w = -1.087 +/- 0.091   (1.0 sigma from -1, below)
  bin 3  z=1.0-2.5 :  w = -1.430 +/- 0.242   (1.8 sigma from -1, below)
TEST 2 — kill tests: Delta chi2 (LambdaCDM -> w0waCDM)
  full: BAO+SN+CMB         Delta chi2 =   7.0   (w0=-0.85, wa=-0.55)
  drop SNe z<0.1           Delta chi2 =   4.7   (w0=-0.80, wa=-0.68)
  drop ALL SNe (BAO+CMB)   Delta chi2 =   6.3   (w0=-0.47, wa=-1.58)
  drop BAO (SN+CMB)        Delta chi2 =   0.8   (w0=-0.89, wa=-0.34)
  wrote prove_binned_w.png

  FACTS:
    * Bin 1 (z<0.4):  w = -0.92 +/- 0.04  -> 2.0 sigma ABOVE -1.
    * Bin 2 (0.4-1):  w = -1.09 +/- 0.09  -> 1.0 sigma below -1.
    * The bins straddle -1 with NO functional form imposed: the crossing is a
      property of the data.
    * Kill tests: the preference survives dropping the low-z SNe (4.7) and even
      ALL SNe (6.3) — but DIES without BAO (0.8): DESI BAO is the load-bearing
      dataset, and the signal lives or dies at DR3.
    * Status: 2.0-sigma-level evidence tonight; 5-sigma requires DESI DR3 +
      LSST-era SNe + full CMB likelihood. This pipeline reruns on those the day
      they drop.

B · Yours not runnable in-browser
This script needs data files, a heavy dependency,
or more time than a browser tab should take.
Download it from /data and run it locally.
Share this claim X Bluesky LinkedIn Reddit HN Email