The Mystery That Launched a Thousand Papers — Solved?
How a 40-Year-Old Puzzle Led Me to a Universal Law of Particle Masses
In a recent physics post, I showed you a single formula that fits every fermion mass ratio in nature using just four numbers. Quarks, electrons, neutrinos — all of them. But I never told you how I knew to look for it.
That’s this story. And it starts with one of the most tantalizing mysteries in all of physics.
A Number That Shouldn’t Exist
In 1981, a Japanese physicist named Yoshio Koide noticed something extraordinary. Take the three charged leptons — the electron, the muon, and the tau. These are the electron and its two heavier cousins: the muon is about 207 times heavier than the electron, and the tau is about 3,477 times heavier. Their masses seem arbitrary, chosen by nature through some inscrutable process.
But Koide found that if you write down this particular combination — using mₑ = 0.511 MeV, mμ = 105.66 MeV, and mτ = 1776.86 MeV:
you get:
Two-thirds. To five decimal places.
Let that sink in. Three masses spanning a factor of 3,477 — masses that appear to follow no pattern whatsoever — produce a ratio that is almost exactly 2/3 when you plug them into this formula.
This is either a cosmic coincidence or a clue. For over forty years, physicists have been unable to determine which.
Why This Drove Physicists Crazy
The Koide relation isn’t like other “near misses” in physics. It’s too precise to ignore and too mysterious to explain. Hundreds of papers have been written about it. Some tried to extend it to quarks — those attempts failed, because if you naively plug quark masses into the same formula, you don’t get anything special. Others tried to derive it from symmetry principles, flavor models, or grand unified theories — but none of these explanations stuck.
The formula just sat there: beautiful, precise, unexplained.
And here’s what made it so maddening: Koide’s formula involves a dimensionless ratio. The masses themselves have units — electron volts, kilograms, whatever convention you prefer. But K is a pure number. If aliens in another galaxy measured the same particles in completely different units, they’d get the same 2/3.
That matters, because dimensionless quantities are nature’s real choices. The absolute mass of the electron depends on our measurement conventions. But a dimensionless ratio like K depends on nothing but the laws of physics themselves. Dimensionless quantities should, in principle, be calculable from first principles.
Which means 2/3 isn’t an accident. It’s a message. The question was: from whom?
Three Generations, No Explanation
Before I tell you what I found, let me set the stage.
The world is made of atoms. Atoms are electrons orbiting a nucleus. The nucleus consists of protons and neutrons. And protons and neutrons are made of quarks — specifically, “up” quarks and “down” quarks.
That’s generation #1: the electron, the up quark, and the down quark. That’s all you need to build every atom in the periodic table, every molecule in your body, every star in the sky.
Yet nature, for reasons no one understands, created two additional generations that are exact copies of these three particles, just heavier. The muon and the tau are copies of the electron. The strange and charm quarks are copies of the up and down. The bottom and top quarks are a third copy.
These heavier copies appear to serve no purpose. They decay almost instantly into their lighter cousins. The top quark — the heaviest of them all, weighing as much as an entire gold atom — survives for about 0.0000000000000000000000005 seconds before it disintegrates. Why does nature bother creating it?
And the mass ratios between generations are bizarre. The top quark is nearly 80,000 times heavier than the up quark. These aren’t small variations — they’re enormous, irregular jumps that nobody can explain.
The Standard Model of particle physics, for all its triumphs, simply takes these masses as inputs. It doesn’t predict them. They’re among the deepest unexplained numbers in all of science.
What I Discovered
I found the reason Koide’s formula works. And it’s not a coincidence — it’s required by a mathematical symmetry rooted in string theory. The paper has been peer-reviewed and published; the math is the math.
Let me explain this at three levels of depth. Pick the one that suits you.
The headline version: There’s a mathematical consistency condition — called modular invariance (the requirement that the physics doesn’t change when you re-parameterize the shape of the extra dimensions, much as Einstein required that physics not depend on your choice of coordinates) — that comes from the way extra dimensions are folded up in string theory. This condition forces the Koide ratio to be exactly 2/3. It’s not optional. It’s not approximate. The symmetry demands it.
Going a bit deeper: In string theory, the extra dimensions beyond our familiar three can be “folded” in specific ways called orbifolds. Think of it like origami — you take a flat sheet and fold it so that certain points are identified with each other, creating corners and symmetry points.
When you fold the extra dimensions this way, the folding creates two natural “building blocks” — two different ways to symmetrically combine the three masses in a given sector. Call them S₁ and S₂. For the mathematically inclined: S₁ is the sum of the masses raised to one power, S₂ is the sum raised to a different power, and those powers are determined by the folding.
Now here’s the key: there is exactly one ratio you can construct from those two building blocks that is (a) symmetric — it doesn’t matter which mass you call “first” — and (b) scale-independent — it doesn’t change if you measure masses in electron volts or kilograms or banana-weights.
Just one. And when you compute it for the electron, muon, and tau: 2/3.
Not because you chose the formula to give 2/3. Because the folding of the extra dimensions only allows one formula, and that formula happens to give 2/3.
The full story in plain English:
If the next few paragraphs make your head spin, skip ahead to “The Quark Surprise” — you won’t lose the thread of the argument.
Here’s how the construction actually works.
In a particular class of string theory constructions called orbifold conformal field theories, the extra dimensions are compactified using discrete symmetries — a ℤₖ group, where k is called the orbifold order. The idea is simple: you rotate the extra dimensions by 360°/k and declare that the rotated configuration is identical to the original. For ℤ₂, you rotate by 180°; do it twice and you’re back where you started. For ℤ₃, you rotate by 120°; three steps to return.
This folding creates “twisted sectors” — particles that are trapped at the symmetry points of the fold, like a crease in a piece of paper. These twisted sectors come with characteristic fractional numbers called “degrees,” one from the left-moving and one from the right-moving vibrations of the string. (A string vibrates in two directions independently — “left-movers” and “right-movers.” The chiral degree comes from one direction alone; the left-right combined degreecombines both directions.) These two degrees give us our two building blocks:
The factor of 2 in the second exponent comes from a basic property of the theory: the total scaling dimension (a number that tells you how a quantity changes when you zoom in or out — essentially, how it “scales” with energy) is twice the chiral dimension, because you have both left-moving and right-moving contributions.
Now, I want a ratio — an observable — that respects three conditions:
Permutation symmetry: It shouldn’t matter which mass I label “1,” “2,” or “3.”
Scale independence (degree zero): If I rescale all masses by the same factor, the ratio shouldn’t change.
Charge quantization: The powers in the ratio should be consistent with the electric charges of the particles, as enforced by modular invariance (the same coordinate-independence condition I described above — it constrains which combinations of charges and orbifold orders are mathematically self-consistent).
It turns out that conditions 1 and 2 together force the ratio to have the form S₁^α / S₂^(2α) for some power α. Why the factor of 2? Because S₁ has homogeneity degree 1/k (each mass is raised to 1/k), and S₂ has degree 1/(2k). For the ratio to be scale-independent — degree zero — the total degree upstairs must equal the total degree downstairs: α/k = β/(2k), which forces β = 2α. And condition 3 — which comes from requiring mathematical consistency of the orbifold — fixesα uniquely: α = kq, where k is the orbifold order and q is the minimal electric charge quantum.
The result is:
This is the master formula. It has no free parameters — once you specify the particle sector (which fixes k and q), the formula is completely determined.
For charged leptons (i.e., the electron, muon, and tau), the orbifold order is k = 1 and the charge quantum is q = 1. Plug those in:
That’s Koide’s formula. And it gives 2/3.
The Quark Surprise
Here’s where it gets really interesting. Half the physics world had tried to extend Koide to quarks and failed. The reason they failed is that they were plugging quark masses into Koide’s formula — the one with square roots. But the orbifold framework says: don’t do that. Quarks have different quantum numbers than leptons, so they live in a different orbifold sector, which means k and q are different, which means the formula itself is different.
For quarks, the orbifold order is k = 2 (a ℤ₂ symmetry) and the charge quantum is q = 1/3 (because the smallest quark charge is 1/3 of the electron’s charge). Plug those into the master formula:
Note: this sums over all six quark masses. And when you evaluate it using the measured masses at the Z-boson scale:
One-half. To four decimal places.
No one had found this before, because no one had the right formula to look for. They kept trying the lepton formula on quarks. But the orbifold says: different charges, different folding, different formula, different answer.
And notice the beautiful structure: the master formula R(k, q) contains both results. When k = 1 and q = 1, you get Koide. When k = 2 and q = 1/3, you get the quark ratio. The framework generates different formulas for different sectors from a single principle. You don’t have to guess the right formula — the mathematics hands it to you.
A note for physicists: the formula uses electric charge |q|, not hypercharge Y — and the absolute value |q|, not the signed charge q. Both choices are surprising. Above the electroweak scale, hypercharge is the “fundamental” quantum number, and electric charge only emerges after symmetry breaking — yet the mass spectrum organizes itself around the broken quantity. Hypercharge does not work. And the sign of the charge, which distinguishes up-type from down-type fermions, is also irrelevant — only the magnitude matters. The same is true of the universal formula from my previous post: replacing |q| with |Y| destroys the fit (because hypercharge no longer distinguishes up-type from down-type quarks), and using signed q instead of |q| also fails. Note, however, that the two formulas use charge differently: the orbifold formula uses the minimum charge quantum of the sector (|q| = 1 for leptons, |q| = 1/3 for all quarks), while the universal formula uses each particle’s own charge (1, 1/3, or 2/3). Yet both insist on |q| over Y. This may be telling us something deep about the relationship between flavor structure and electroweak symmetry breaking.
It’s Not a Coincidence
Two natural questions: Is 1/2 really exact? And does it stay 1/2 at different energy scales?
On the first: the value Rq = 0.4995 is consistent with 1/2 within experimental uncertainties. The masses of light quarks (up, down, strange) are notoriously hard to measure — their uncertainties are 10–20%. The fact that six masses with these large individual uncertainties conspire to produce a ratio within 0.1% of 1/2 is remarkable.
On the second: in quantum field theory, particle masses “run” — they change depending on the energy scale at which you probe them. The individual quark masses change enormously between the electroweak scale and 10¹⁴ GeV. But I showed that Rq barely budges. At one loop (the first level of quantum correction, where you account for virtual particles popping in and out of the vacuum once), it’s exactly invariant — due to a cancellation involving the “gauge group structure” (don’t worry about what that means; the point is that it cancels exactly, not approximately). At two loops (the next level of correction, which is far more intricate), it shifts by only a few parts in a thousand over twelve orders of magnitude in energy.
This means the relation is fundamental, not accidental. An accident would be destroyed by quantum corrections. A consequence of deep symmetry survives them — and that’s exactly what happens.
The high-scale prediction is: Rq(10¹⁴ GeV) = 0.500 ± 0.002. Upcoming lattice QCD calculations (a technique that simulates quark interactions on a discrete grid of spacetime points using supercomputers, allowing physicists to extract quark masses from first principles) should be able to test this.
The Moment Everything Changed
So here’s where I stood: I had explained a 40-year mystery (Koide’s 2/3) and discovered something new (the quark ratio of 1/2), both flowing from the same mathematical structure with zero free parameters.
And then the thought hit me.
If dimensionless mass ratios in the lepton sector are governed by a mathematical symmetry, and dimensionless mass ratios in the quark sector are also governed by the same symmetry... then there’s hidden structure in the fermion mass spectrum that no one has been fully exploiting.
And if there’s hidden structure, there should be a universal expression for all fermion mass ratios — not a separate formula for each sector, but one single relation that captures the whole pattern. A relation built from the most basic labels that distinguish particles: the generation number (1, 2, or 3) and the electric charge.
After all, those were the only ingredients that went into the orbifold construction: k depends on the charge family, and the formula is symmetric across generations. The deeper pattern should depend on nothing else.
So I went looking for it. And as readers of my previous post already know, I found it.
(And before you say “overfitting” — I tested that. I ran all 120 possible assignments of generation labels to masses; only the physical one fits. I scanned all 720 possible ratio bases; only one converges. Alternative functional forms are decisively ruled out by information criteria. The details are in the previous post and in the paper.)
The Larger Picture
Let me step back and tell you what I think this means.
For forty years, Koide’s formula was treated as a curiosity — a beautiful but isolated numerical coincidence. Nobody could explain it, nobody could extend it, and gradually most physicists filed it away as “probably a coincidence.”
It’s not a coincidence. It’s the tip of an iceberg.
The reason no one could extend Koide to quarks is that they were looking for the same formula in a different sector. But the orbifold framework says: the formula should be different for quarks, because quarks carry different charges and live in a different part of the geometry. What’s universal is not the formula but the principle — modular covariance (the requirement that the mass functional transforms in a controlled, predictable way under re-parameterizations of the extra-dimensional geometry, rather than being completely invariant) — that generates the right formula for each sector.
And once you see that the principle is universal, you realize the mass spectrum is trying to tell us something. The twelve fermion masses aren’t arbitrary constants. They’re constrained by deep mathematical structure — structure that might, ultimately, come from the shape of the extra dimensions in which we live.
The paper is published in Modern Physics Letters A and available here: https://www.worldscientific.com/doi/10.1142/S0217732326500732
The universal pattern paper is in the European Physical Journal C: https://link.springer.com/article/10.1140/epjc/s10052-025-14771-0
What’s Next
The universal formula from the previous post has four parameters:
where d is the generation index, q is the electric charge, and the fitted values are κ ≈ 2.3, γ ≈ 1.1, ζ₁ ≈ 1.17, ζ₂ ≈ −0.8 — all of order 1.
I’m now working on deriving those four parameters from first principles. If that succeeds, zero free parameters remain — and the twelve-order-of-magnitude fermion mass spectrum would be entirely determined by mathematical structure.
Stay tuned.
If you enjoyed this, you might enjoy my book The Science of Free Will. The fermion mass puzzle and free will might seem unrelated, but they share a deep question: can a system governed by rigid mathematical laws produce outcomes that are, in any meaningful sense, unpredictable? The answer — computational irreducibility — changes how you think about everything from particle physics to human choice.
Nice!