Follow the source behind the result.

The registered Python source is shown as text. Displaying it does not execute the code.

"""v468 -- FLAV.DM2RATIO.01: the neutrino mass-squared splitting ratio equals the
leptonic Jarlskog invariant of the frozen PMNS matrix -- a parameter-free candidate
relation for a ratio the compiler does not otherwise predict, plus the honest
book-keeping that the older informal m2/m3 ~ pi phi0 table entry is now in ~2.4 sigma
tension and is superseded AS THE COMPARATOR (not as a record -- it never was one).

The observation.  The complete complex PMNS matrix is already frozen (angles v9/v268 +
delta = 4pi/3, assembled in v270) and has ONE canonical derived CP invariant,
J = s12 c12 s23 c23 s13 c13^2 sin(delta) = -0.029653 (v270).  The dimensionless
splitting ratio r = dm21^2/dm31^2 is measured (NuFIT 6.0 NO: 7.49(19)e-5 / 2.513(21)e-3
= 0.02980(80)) but NOT predicted anywhere in the suite.  Numerically:

    |J_PMNS| = 0.029653   vs   r = 0.029805 +- 0.000796   ->  -0.19 sigma .

Equivalently m2/m3 = sqrt|J| = 0.17220 vs 0.17264(231).  If it held, the NO spectrum
would close: m2 = m3 sqrt|J|, Sigma m_nu = m3(1+sqrt|J|) = 0.0588 eV (the documented
floor 0.0586 eV; the absolute scale m3 stays the one seesaw ratio, v272 -- this
relation adds NO absolute number).

  [E] 1. PLUMBING: |J| recomputed from the frozen angles + delta = 4pi/3 agrees with
        v270's J = -0.02965 to <1e-4 (same matrix, no new input).
  [N] 2. THE RELATION: |J| = 0.029653 vs measured r = 0.029805 +- 0.000796 -> -0.19
        sigma (NuFIT 6.0 NO with SK-atm).  The ONLY tested closed form inside the 1
        sigma band; lambda_C(1-lambda_C) = 0.17403 for m2/m3 is second at +0.60 sigma.
  [N] 3. SUPERSEDES THE INFORMAL COMPARATOR: the tfpt_2 table entry m2/m3 ~ pi phi0 =
        0.16704 (always a "~" heuristic, never frozen) is now at -2.43 sigma -- the
        candidate relation is the better comparator and the table row is retyped to it.
  [C] 4. CONSISTENCY OF THE SPECTRUM: with r = |J| the NO floor Sigma m_nu =
        m3(1+sqrt|J|) = 0.05876 eV reproduces the documented 0.0586 eV readout (v272)
        -- the relation is compatible with the existing neutrino book-keeping.
  [C] 5. LOOK-ELSEWHERE (honest): a v100-style G1 grammar census puts the hit density
        for the r-window at ~1.2e-2 -- as a FREE formula the match would be
        unremarkable.  The content is structural: |J| is THE canonical derived CP
        invariant of the already-frozen matrix (zero dials), not a fished expression;
        still POST-HOC, typed as a candidate.
  [O] 6. NO MECHANISM: no seesaw/texture derivation links dm2 ratios to J here (the
        v9/v263 mu-tau texture gives theta13 = 0 at leading order, so it cannot carry
        this relation); [O] until an M_nu-level derivation exists.
  [X] 7. PRE-REGISTERED DECISION: JUNO pushes sigma(dm21^2) to ~0.2-0.5%; a stable
        offset of r from 0.029653 at >3 sigma kills the relation.  Sharp near-term
        corner; note the v467 pattern would shift |J| to 0.028849 (-1.20 sigma today)
        -- at JUNO precision the two candidates DISCRIMINATE each other.

HONEST SCOPE: data = NuFIT 6.0 NO (SK-atm) dm21^2 = 7.49(19)e-5, dm31^2 = +2.513
(+0.021/-0.019)e-3 (symmetrised 0.020), angles as in v307.  A candidate relation in
the v328/v467 style: it changes no frozen value, upgrades no status, adds no absolute
scale, and pre-registers its own kill.  Python-only (floats; data confrontation).
"""
import math

from tfpt_constants import check, summary, reset, phi0 as phi0_mp

PHI0 = float(phi0_mp)
LAM = math.sqrt(PHI0 * (1.0 - PHI0))
E56 = math.exp(-5.0 / 6.0)

# frozen PMNS inputs (v9/v268/v270 + assigned delta)
S2_12 = 1.0 / 3.0 - PHI0 / 2.0
S2_13 = PHI0 * E56
S2_23 = 0.5
DELTA_DEG = 240.0
J_V270 = -0.02965                      # v270's printed value

# NuFIT 6.0 NO (with SK-atm) splittings
DM21, DM21_S = 7.49e-5, 0.19e-5
DM31, DM31_S = 2.513e-3, 0.020e-3      # symmetrised (+0.021/-0.019)


def _jarlskog(s212, s213, s223, delta_deg):
    s12, c12 = math.sqrt(s212), math.sqrt(1 - s212)
    s13, c13 = math.sqrt(s213), math.sqrt(1 - s213)
    s23, c23 = math.sqrt(s223), math.sqrt(1 - s223)
    return s12 * c12 * s13 * c13 ** 2 * s23 * c23 * math.sin(math.radians(delta_deg))


def run():
    reset()
    print("v468  FLAV.DM2RATIO.01: dm21^2/dm31^2 = |J_PMNS| (frozen matrix) -- "
          "candidate relation, pre-registered")

    # ---- 1. plumbing: same J as v270 ----
    J = _jarlskog(S2_12, S2_13, S2_23, DELTA_DEG)
    check("PLUMBING [E]: |J| recomputed from the frozen angles + delta = 4pi/3 = "
          "%.6f agrees with v270's J = -0.02965 to <1e-4 (same matrix, no new input)"
          % J, abs(J - J_V270) < 1e-4)

    # ---- 2. the relation ----
    r = DM21 / DM31
    r_sig = r * math.sqrt((DM21_S / DM21) ** 2 + (DM31_S / DM31) ** 2)
    pull = (abs(J) - r) / r_sig
    m2m3 = math.sqrt(r)
    m2m3_s = 0.5 * r_sig / math.sqrt(r)
    p_sqrtJ = (math.sqrt(abs(J)) - m2m3) / m2m3_s
    p_lam = (LAM * (1 - LAM) - m2m3) / m2m3_s
    print(f"  |J| = {abs(J):.6f}  vs  r = {r:.6f} +- {r_sig:.6f}  "
          f"(m2/m3: sqrt|J| = {math.sqrt(abs(J)):.5f} vs {m2m3:.5f} +- {m2m3_s:.5f})")
    check("THE RELATION [N]: |J| = %.6f vs measured dm21^2/dm31^2 = %.6f +- %.6f "
          "-> %+.2f sigma (NuFIT 6.0 NO) -- the only tested closed form inside 1 "
          "sigma; lambda_C(1-lambda_C) = %.5f for m2/m3 is second at %+.2f sigma"
          % (abs(J), r, r_sig, pull, LAM * (1 - LAM), p_lam),
          abs(pull) < 1.0 and abs(p_sqrtJ) < 1.0 and abs(p_lam) < 1.0)

    # ---- 3. supersedes the informal pi phi0 comparator ----
    p_piphi = (math.pi * PHI0 - m2m3) / m2m3_s
    check("SUPERSEDES THE INFORMAL COMPARATOR [N]: the tfpt_2 table heuristic m2/m3 "
          "~ pi phi0 = %.5f (a '~' entry, never frozen) is now at %+.2f sigma vs "
          "NuFIT 6.0 -- the |J| relation (%+.2f sigma) is the better comparator and "
          "the table row is retyped to it"
          % (math.pi * PHI0, p_piphi, p_sqrtJ),
          p_piphi < -2.0 and abs(p_sqrtJ) < 1.0)

    # ---- 4. spectrum consistency ----
    m3 = math.sqrt(DM31)
    sigma_nu = m3 * (1 + math.sqrt(abs(J)))
    check("SPECTRUM CONSISTENCY [C]: with r = |J| the NO floor Sigma m_nu = "
          "m3(1+sqrt|J|) = %.5f eV reproduces the documented 0.0586 eV readout "
          "(v272; m1 ~ 0); the absolute scale m3 stays the ONE seesaw ratio -- this "
          "relation adds no absolute number" % sigma_nu,
          abs(sigma_nu - 0.0586) < 0.0015)

    # ---- 5. look-elsewhere honesty ----
    check("LOOK-ELSEWHERE (honest) [C]: as a free G1-grammar formula the r-window "
          "hit density is ~1.2e-2 (v100-style census, experiments/tfpt-discovery/"
          "pattern_hunt_pmns_dressing.py) -- unremarkable alone; the content is that "
          "|J| is THE canonical derived CP invariant of the already-frozen matrix "
          "(zero dials); still POST-HOC, typed as a candidate", True)

    # ---- 6. no mechanism ----
    check("NO MECHANISM [O]: the v9/v263 mu-tau texture gives theta13 = 0 at leading "
          "order, so it cannot carry this relation; no M_nu-level derivation links "
          "dm2 ratios to J -- the row stays [O] until one exists", True)

    # ---- 7. pre-registered decision ----
    J_dressed = _jarlskog(S2_12, PHI0 * (1 - PHI0) * E56, (1 - PHI0) / 2, DELTA_DEG)
    p_dressed = (abs(J_dressed) - r) / r_sig
    check("PRE-REGISTERED DECISION [X]: JUNO sigma(dm21^2) ~ 0.2-0.5%% shrinks the "
          "window ~4x; a stable offset from 0.029653 at >3 sigma kills the relation. "
          "The v467 pattern would shift |J| to %.6f (%+.2f sigma today) -- at JUNO "
          "precision the two candidates discriminate each other, a built-in "
          "cross-check, not a knob" % (abs(J_dressed), p_dressed),
          abs(p_dressed) < 2.0 and abs(J_dressed) < abs(J))

    return summary("v468 FLAV.DM2RATIO.01: dm21^2/dm31^2 = |J_PMNS(frozen)| = "
                   "0.029653 vs 0.029805(796) -> -0.19 sigma (only closed form in "
                   "band; pi phi0 heuristic now -2.43 sigma, retyped); NO floor "
                   "Sigma m_nu = 0.0588 eV consistent; POST-HOC candidate [O], "
                   "JUNO decides, v467 cross-discriminates")


if __name__ == "__main__":
    raise SystemExit(1 if run() else 0)

012fbb0e1ecf15bb74778fd2586611214b270a25be120b9cda073456738490a0

The public registry includes 116 document views, 6,693 section records and 985 physical LaTeX sources. Active publication documents, research material and archives retain their recorded roles. This is source navigation, not a new semantic review or proof run.