Papex

Open problem · Physics · since 1981 · Open

Koide formula

Why do the masses of the electron, muon and tau fit Koide’s simple formula so precisely?

Summary

In 1981 Yoshio Koide noticed that a simple ratio built from the three charged lepton masses equals 2/3 to within measurement error; whether this reflects deeper physics or is a numerical coincidence is unknown.

Why it is open

The relation states that the sum of the three charged-lepton masses, divided by the square of the sum of their square roots, equals 2/3 [5]. This ratio can only lie between 1/3, when the three masses are equal, and 1, when one mass dominates; Zhi-zhong Xing and He Zhang note that it "might be a puzzle" that the charged leptons sit exactly at the mean of the two [5]. Robert Foot showed that the relation is equivalent to an angle of exactly 45 degrees between the vector of the square roots of the masses and the vector (1, 1, 1) [2].

The Standard Model offers no prediction to compare with: the charged-fermion masses come from Yukawa couplings (the strengths with which each fermion couples to the Higgs field), which it leaves "completely undetermined" [9]. Why there are three generations at all is a separate problem, the generations of matter.

The relation holds for pole masses, the physical masses that experiments measure [5, 7]. Theories at high energies predict running masses, which quantum corrections shift and which change with the energy scale [5, 7]. Foot argued in 1994 that the same relation cannot hold for the bare mass parameters "because the electromagnetic corrections are too big", which makes it "difficult to understand" from any simple gauge model [2]. In 2009 Yukinari Sumino wrote that no realistic model or mechanism had been found that predicts the formula within the required accuracy, and that the most serious problem is the correction from quantum electrodynamics (QED) [7].

No body sets a bar for a resolution. The sources point to two outcomes: a mechanism that gives the relation for pole masses at the measured precision, including a reason for the cancellation of the QED correction [7]; or a failure of the relation as the tau mass is measured more precisely, which Foot proposed watching because its success may be "just a numerical coincidence" [2].

What is known

Origin (1981–1983). Alejandro Rivero and Andre Gsponer, in a historical review posted as a preprint in 2005, write that Koide came across the relation at the end of 1981 while working on composite models of quarks and leptons [3]. It predicted a tau mass of 1.777 GeV, then more than two standard deviations from the measured 1.7842 GeV, so in 1983 Koide published a phenomenological version with a correction term fitted to the measurement [3]. That 1983 paper, on a three-family model, reported "an excellent prediction" of 1.7866 GeV from the electron and muon masses alone [1]. Koide later wrote that he derived the relation, proposed in 1982, from a composite model, without considering permutation symmetry [6].

The tau mass revised (1992–1994). Foot reports that Koide had pointed out the agreement of new tau mass measurements by the ARGUS, BES and CLEO collaborations with the prediction of 1776.97 MeV [2]. With the revised mass, Rivero and Gsponer write, the correction term "becomes plainly zero" [3].

Precision (2006–2026). With the data of 2006, Xing and Zhang found the ratio within −0.00001 and +0.00002 of 2/3 [5]. Xing's 2020 review, using pole masses that include a tau mass of 1776.86 ± 0.12 MeV, finds the relation satisfied "up to the accuracy of O(10⁻⁵)" and adds: "Whether such an amazing equality has a deeper meaning remains to be seen" [9]. In an October 2026 preprint, Alessandro Strumia writes that with the 2024 masses of the Particle Data Group the relation "holds within 1σ, with a relative accuracy of order 10⁻⁵" [11].

Running masses (2006–2009). Xing and Zhang found that running masses "fail in satisfying" the relation, although the discrepancy is "only about 0.2%" at the mass of the Z boson, and that the ratio stays nearly stable from there up to about 10¹⁴ GeV [5]. Nan Li and Bo-Qiang Ma, studying the same running, concluded that the relation is "independent (or insensitive)" of energy scale from low to extremely high energies; their table gives the ratio at 91 GeV as 1.001881 times the Koide value [4]. Sumino stated the consequence: if running masses satisfied the relation far above the W boson mass, the pole masses would miss it by about 120 times the experimental error, so any high-scale origin needs a correction that cancels the QED one [7].

Other fermions (2005–2011). Rivero and Gsponer note that the down-type quark masses are approximated "well enough" by the formula and the up-type masses less well [3]. At the Z mass, Xing and Zhang find the ratio near 0.89 for the up-type and 0.74 for the down-type quarks [5]. Werner Rodejohann and He Zhang showed that the light neutrino masses cannot obey the relation, but that with the seesaw mechanism it could hold for the masses in the Dirac or heavy Majorana mass matrices [8].

Recent claims. In 2026 Rivero reported in Physics Letters B an "inverse" rule of Koide type for the down-type quarks, built from the inverse square roots of their masses [10]. With central mass values its ratio reaches exactly 2/3 near 280 TeV under Standard Model running, and it stays within one standard deviation of 2/3 along the whole run [10]. The paper declares that Claude and GPT 5.4 were used for "algebraic exploration, numerical checks, and minor drafting assistance", and that the relation became apparent during a numerical exploration launched by Claude Opus 4.6 [10]. Rivero cautions that as the required precision is relaxed, "it is really not so complicated to find random masses fitting into them" [10]. INSPIRE listed no papers citing it on 16 October 2026 [10].

Main approaches

Flavour symmetry in the scalar sector. The lepton masses come from the vacuum values of three flavour scalars, and a symmetric potential makes those values satisfy the Koide condition [6]. Strongest evidence: Koide shows that a potential invariant under S3, the group of permutations of three objects, gives the required relation among the vacuum values whatever the value of its key coupling [6]. Main difficulty: every such model, Sumino wrote, "requires either absence or strong suppression of some of the terms in the potential" without justification, and before his work no mechanism had been proposed for the QED correction [7].

Family gauge symmetry. Sumino proposed a U(3) gauge symmetry acting on the three families, whose radiative correction cancels the QED correction to the formula, in a scenario where it unifies with the weak SU(2) symmetry at 10²–10³ TeV [7]. Strongest evidence: in this effective theory the cancellation follows from the structure of the corrections, and the scenario predicts lepton-flavour-violating processes at that scale [7]. Main difficulty: Sumino states that some accidental factor or parameter tuning cannot be avoided, that extra fermions are needed to cancel anomalies, and that the experimental bound on the decay of the long-lived neutral kaon to an electron and a muon "may be marginally in conflict" with his estimated unification scale, depending on assumptions about the quarks [7].

Mass-matrix textures. Lepton mass matrices are chosen so that their structure enforces the relation and also fits neutrino mixing [11]. Strongest evidence: Strumia's textures realise the relation and fit the measured lepton mixing angles, with predictions for the CP-violating phases [11]. Main difficulty: he concludes that the textures realising the relation "are contrived and lack a compelling symmetry interpretation", and that their numerical success "was achieved after exploring many possibilities" [11].

Coincidence. The relation may be "of no fundamental significance", its success "just a numerical coincidence", as Foot allowed in 1994 [2]. Strongest evidence: no realistic mechanism was known in 2009 [7], and close matches are easy to find: Rivero and Gsponer note a ratio from their quark-mass table "quite precisely equal to π/2", which "illustrates how easily one can get a good agreement with a beautiful number" [3]. Main difficulty: the relation has held at the level of 10⁻⁵ in assessments from 2006, 2020 and 2026 [5, 9, 11], a precision Xing and Zhang call "so amazing that there might exist some underlying physics" [5].

Open sub-questions

  1. The next tau mass. The electron and muon masses fix a Koide prediction of about 1776.97 MeV for the tau [11]. Does the relation still hold within one standard deviation once the tau mass is measured with a smaller uncertainty than in the 2024 averages [11]?
  2. Family gauge bosons. In Sumino's U(3) scenario, which rates of lepton-flavour-violating decays follow for family gauge bosons at 10²–10³ TeV, and how much of that range does the kaon bound already exclude [7]?
  3. Where the inverse rule is exact. Does the scale at which Rivero's inverse down-quark ratio crosses 2/3, 280 TeV with one running code and 10.18 TeV with interpolated tables, converge as the down and strange quark masses become more precise [10]?
  4. Chance level. For masses drawn from a log-uniform distribution, how often does a triple satisfy a direct or inverse Koide-type rule at the precision seen for the charged leptons, once the number of possible rules and triples is counted [3, 10]?
  5. Seesaw neutrinos. If the relation holds for the heavy Majorana masses of the seesaw mechanism, which neutrinoless double-beta decay rates follow [8]? See also Neutrino mass hierarchy.

Work on this in Papex

If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.

Key references

  1. Koide, Y. (1983). New view of quark and lepton mass hierarchy. Physical Review D 28(1), 252–254. 10.1103/PhysRevD.28.252 · Checked 16 Oct 2026
  2. Foot, R. (1994). A note on Koide's lepton mass relation. Preprint, arXiv:hep-ph/9402242 (version 1, 7 February 1994), not peer reviewed. Source · Checked 16 Oct 2026
  3. Rivero, A. & Gsponer, A. (2005). The strange formula of Dr. Koide. Preprint, arXiv:hep-ph/0505220 (version 1, 25 May 2005), not peer reviewed. Source · Checked 16 Oct 2026
  4. Li, N. & Ma, B.-Q. (2006). Energy scale independence of Koide's relation for quark and lepton masses. Physical Review D 73(1), 013009. 10.1103/PhysRevD.73.013009 · Checked 16 Oct 2026
  5. Xing, Z.-z. & Zhang, H. (2006). On the Koide-like relations for the running masses of charged leptons, neutrinos and quarks. Physics Letters B 635(2–3), 107–111. 10.1016/j.physletb.2006.02.051 · Checked 16 Oct 2026
  6. Koide, Y. (2007). Charged lepton mass formula: development and prospect. International Journal of Modern Physics E 16(5), 1417–1426. 10.1142/S0218301307006770 · Checked 16 Oct 2026
  7. Sumino, Y. (2009). Family gauge symmetry and Koide's mass formula. Physics Letters B 671(4–5), 477–480. 10.1016/j.physletb.2008.12.060 · Checked 16 Oct 2026
  8. Rodejohann, W. & Zhang, H. (2011). Extension of an empirical charged lepton mass relation to the neutrino sector. Physics Letters B 698(2), 152–156. 10.1016/j.physletb.2011.03.007 · Checked 16 Oct 2026
  9. Xing, Z.-z. (2020). Flavor structures of charged fermions and massive neutrinos. Physics Reports 854, 1–147. 10.1016/j.physrep.2020.02.001 · Checked 16 Oct 2026
  10. Rivero, A. (2026). New sum rules of the Koide type. Physics Letters B 877, 140510. 10.1016/j.physletb.2026.140510 · Checked 16 Oct 2026
  11. Strumia, A. (2026). A texture of leptons for the Koide and Alikhanov relations. Preprint, arXiv:2610.10696 (version 1, 7 October 2026), not peer reviewed. Source · Checked 16 Oct 2026

Page reviewed 16 Oct 2026

Source

Open, as listed by Wikipedia · Unsolved problems in physics. Checked 08 Oct 2026.

The status is the source’s. Papex does not decide whether a problem is solved.

Records on this problem

No public record addresses this problem yet.

More in Physics