The sharpest empirical handle on the vacuum: w = −1 is an identity for a true vacuum energy (Lorentz invariance forces T_μν = −ρ g_μν), not a fit parameter — so observation can test it. From this registry's corrected CPL fit (w₀ = −0.849, wₐ = −0.548), w crosses −1 at z = 0.380 and is phantom above it, at Δχ² = 6.97 on 2 dof = 2.16σ. That is NOT a detection: 24% of the way to 5σ, which needs σ(w₀) ≤ 0.030.

OSED FORM vs NUMERICAL rho_DE: max relative disagreement
       best fit: 3.094e-14 (GL in ln a), 1.861e-15 (composite GL in a),
       2.613e-13 (RK4 on the ODE).  Arithmetic validation only -- carries
       ZERO evidence weight about nature.

  [H2] PERCENT CHANGE IN rho_DE, z = 1 -> today:
       best fit (w0=-0.836, wa=-0.609): -1.178 %   [rho ratio 0.988218]
       LCDM / true vacuum:            +0.000 %   EXACTLY, by construction
       Peak excursion: +7.589 % at z = 0.369 (the w=-1 crossing).
       DO NOT quote the z=1 number alone -- it is small only because it
       sits where the rising and falling branches cancel.  Largest
       |drift| in the z <= 2.5 data window: 30.75 % at z = 2.50; at the
       window edge z = 2.5 it is -30.75 %.
       NO ERROR BAR: no covariance was supplied.  PART F shows this
       number's SIGN is not established over a plausible parameter grid.

  [H3] w = -1 CROSSING REDSHIFT (best fit): z_cross = 0.3685 (a = 0.730706).
       Registry reports z ~ 0.37; verified from the fitted parameters.
       LCDM: no crossing -- w is pinned at -1 identically.
       Phantom (w < -1) for ALL z > 0.369.

  [H4] rho_DE(a)/rho_DE(1) as a -> 0 (best fit): -> 0, from above.
       Exponent on a is +1.3350 > 0; exp() factor -> 6.2152 (finite).
       STRICTLY POSITIVE for all a > 0; NO sign change anywhere in
       0 < a <= 1 -- but CPL cannot produce one for ANY parameters, so
       that is a fact about the parameterization, not a measurement.
       Not an energy floor: dark energy switched on.

  [H5] EVIDENCE: Delta chi^2 = 8.2 (2 extra parameters) = 2.40 sigma.
       3 sigma needs Delta chi^2 = 11.83  (have 69%, short by 3.63)
       5 sigma needs Delta chi^2 = 28.74  (have 29%, short by 20.54)
       NOT A DETECTION.  Data are CONSISTENT with a constant vacuum and
       MILDLY PREFER evolution.  Honest label: PROVISIONAL.

  [H6] REQUIRED sigma(w0) FOR A 5-SIGMA STATEMENT at the current central
       value: |w0+1|/5 = 0.164000/5 = 0.032800.
       (3 sigma would need 0.054667.)  Rough current-scale estimate 0.068446
       -> factor 2.09 tighter, ~4.4 x the effective data volume.
       CONDITIONAL on the central value not moving toward -1.  If
       w0 = -1 is the truth, no precision ever gets to 5 sigma.

==============================================================================
INTERPRETATION (honesty rules apply)
==============================================================================
  THIS IS THE ONE PLACE IN THE WHOLE VACUUM-ENERGY QUESTION WHERE DATA,
  NOT PHILOSOPHY, DOES THE WORK.

  Tests 13 and 14 in this wave could not be settled by computation
  because the obstacle there is conceptual: nobody knows whether
  zero-point energy should gravitate, and no amount of arithmetic
  decides it.  The w = -1 question is different in kind.  A vacuum
  energy has w = -1 exactly and forever -- not approximately, not on
  average, not as a best fit.  Lorentz invariance of the vacuum state
  makes it an identity.  So w is a knob that observation can turn, and
  it is turning.

  WHAT THE DATA CURRENTLY SAY, in one sentence: consistent with
  constant, mildly prefers drift, not a detection.

  Quantitatively, from the registry's own fitted values:
    * The measured w(z) crosses -1 near z = 0.37 and is PHANTOM above
      it -- behaviour a vacuum energy is structurally incapable of.
    * The implied density drift is modest but NOT uniformly tiny, and
      the tempting one-line gloss is wrong.  rho_DE rises to ~8% above
      today's value at the crossing, is back within ~1% of today at
      z = 1 (a coincidental cancellation, not a smallness result), and
      is ~31% BELOW today's value by z ~ 2.5 at the edge of the data.
      So: a few-percent wiggle for z < 1, a tens-of-percent decline
      across the full window.  Not a dramatic switch-on; not a wiggle
      either.
    * Delta chi^2 = 8.2 for two extra parameters is 2.40 sigma.  Three
      sigma needs 11.83; five needs 28.74.  We have 69% of the way to
      the weaker of those two bars.  That is a preference.
    * Without the fit covariance -- which the registry did not supply --
      even the SIGN of the z=1 drift is not established across a
      plausible parameter range.  This is the largest single caveat on
      the page and it is a caveat about bookkeeping, not about nature.

  THE CONNECTION, STATED AND NOT OVERCLAIMED.  IF w != -1 is confirmed
  by Euclid, DESI DR3 or LSST, then dark energy is not a fixed energy
  floor and the constant-Lambda interpretation FAILS -- regardless of
  how the cosmological constant problem is resolved.  That is a real
  and unusual logical structure: a measurement that would falsify the
  vacuum reading without needing to understand it first.  Reaching 5
  sigma at today's central value needs sigma(w0) ~ 0.033 -- roughly
  twice today's precision, on a back-of-envelope inversion that PART F
  labels as sloppy.  Whether that happens depends on the central value
  staying put, which is exactly the part nobody can forecast.

  WHAT THIS DOES NOT DO -- SAID EXPLICITLY.
    * NOTHING HERE SOLVES THE COSMOLOGICAL CONSTANT PROBLEM.  Not
      partially, not in principle.  The magnitude discrepancy between
      predicted vacuum energy density and the observed value is ~10^60
      with a TeV cutoff and ~10^120 with a Planck cutoff -- and quoting
      either number without its cutoff assumption is meaningless,
      because the 'prediction' is not the output of any complete
      theory.  A tens-of-percent drift in rho_DE is orthogonal to a
      60-to-120-ORDER-OF-MAGNITUDE mismatch.  Confirming w != -1 changes
      WHICH problem we have, not answer this one.
    * NO NEW FIT IS CLAIMED.  Four numbers (w0, wa, Delta chi^2, and
      the crossing redshift) came in from claim USR-2026-0001.  No
      likelihood was evaluated, no data were read, no chain was run.
      Everything printed above is an exact consequence of those four
      inputs plus textbook cosmology.  REPRODUCED, NOT DISCOVERED.
    * CPL IS NOT A THEORY.  It is a two-parameter Taylor expansion in
      (1-a).  Its w=-1 crossing, its positivity, and its a -> 0
      behaviour are all partly artifacts of that functional form.  The
      strongest honest version of this test's conclusion is a
      CONDITIONAL, and the condition is 'if CPL is the right shape'.
    * THE PROJECT'S OWN AUDIT SAYS 3-4 SIGMA, and published DESI DR2
      analyses span ~2.8 to ~4.2 sigma depending on the supernova
      compilation.  That spread is dataset choice, not statistics.  A
      result whose significance depends on which SN sample you pick is
      by definition not yet a measurement of nature.

  BOTTOM LINE: the vacuum's constancy is TESTABLE, is BEING TESTED, and
  currently PASSES -- weakly, with a mild and unstable pull away from
  -1 that has not earned the word 'detection'.  Provisional.
==============================================================================

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.