pdghist — predictions

A small playground that overlays a couple of theoretical sum-rule predictions on the PDG history. The motivating articles by Alejandro Rivero are New sum rules of the Koide type (Phys. Lett. B 877, 140510, 2026) and the Casimir-quartic / Poincaré manuscript.

« back to the main pdghist page — many of these predictions also appear as dashed pink reference lines on the per-quantity cards there.

Koide tuples

The original Koide formula K = (m₁+m₂+m₃) / (√m₁+√m₂+√m₃)² = 2/3 for charged leptons inspires a whole family of triplet sum rules. Some require taking the negative square root for one of the masses (those entries are marked − below). The down-quark inverse tuple uses K⁻¹ = (1/m₁+1/m₂+1/m₃)/(1/√m₁+1/√m₂+1/√m₃)² = 2/3. Source: Rivero, Phys. Lett. B 877 (2026) 140510.

tuple type predicted K K from PDG 2025 deviation postdicted note (hover for citation)

Square-root sum rule (Eq. 10 of the paper): √m_b = √3·√m_s + √m_c → predicts m_b = 4184.5 MeV (PDG 2025: 4183 ± 7 MeV, match to 0.04%). Drawn as a pink dashed line on the b-quark mass card on the main page.

The down-quark inverse Koide tuple hits the fixed point K⁻¹ = 2/3 exactly at Q ≈ 280 TeV under SM RG running (Martin-Robertson SMDR, 5-loop). At Q = M_Z the value is K⁻¹ = 0.66750 (+0.13% from 2/3); at Planck scale, 0.66539 (−0.19%).

K(year) per tuple

Computed live from the PDG mass history. Pink dashed line is the prediction K = 2/3 = 0.66667.

Poincaré–Casimir quartic spectrum

Identifying the four electroweak masses with a signed quartet of Poincaré representations indexed by spin s and parity, with mass-squared values r² M_Z². The four predicted ratios r come out of a single dimensionless construction and depend on no free parameter beyond M_Z as the input scale. Source: Rivero, Casimir-quartic / Poincaré manuscript.

slot identified with r = ratio to M_Z tree prediction (GeV) with 3/8 correction (GeV) PDG 2025 (GeV)

The fourth column (3/8 correction) applies ε = (C_F/C_A)(M_Z² − M_W²) = (3/8)·(91.188² − 80.369²) ≈ 696 GeV² to the negative-sector slots with the spin-grading sign factor (−1)^(2s+1) (Eq. 33 of the manuscript). The positive slots are unchanged. With this single correction the s = ½ negative slot lands at 125.20 GeV, the s = 1 negative slot at 174.17 GeV, and the s = 3/2 positive slot stays at 96.54 GeV. All four electroweak masses are then reproduced within experimental error.

Equivalent on-shell weak mixing angle: s²_dV = 1 − (β_{1/2}/β_1)² = 0.22310132… (the pink dashed line on the on-shell sin²θ_W card on the main page).

Same construction, read as a one-parameter over-determined fit, gives a Poincaré rest mass m ≈ 106.57 GeV: m_W = M_W/√((√57−3)/8) = 106.5705 GeV, m_Z = M_Z/√(√3−1) = 106.5779 GeV (two agreeing-determinations to a part in 10⁴, reproducing the m ≈ 106.5 GeV Rivero noticed in his 2004 posting).

The construction also has a positive s = 3/2 slot at 96.54 GeV, coincident with the LEP/CMS 95-96 GeV anomaly. No Standard Model particle is currently identified at this mass.

How to read the deviations

For the W mass slot the +0.4σ agreement is striking because the construction has only one input (M_Z), so the W mass is a pure prediction with no fitted parameter — competitive with the global SM electroweak fit's ±5 MeV prediction (80.359 GeV).

The Higgs and the Fermi VEV deviate at the percent level. The paper interprets these as the tree-level signal to be perturbed by O(α) corrections; the splitting structure of the signed negative-sector eigenvalues then naturally produces values close to the observed m_H and v/√2.