Is the Spectrum a Fit?

The escape hatch in most number games is a tunable scale: fit one continuous knob per sector and the integers only have to get close. IDWT does not have that knob. Each sector's mass scale is fixed by the electron mass and the composite \(n_s=4\) through the coupling chain — the same two inputs everywhere, not refit sector by sector. Once those are set, a particle's mass is \(S(n,d)\) times a number the theory already owns, and the only remaining freedom is the choice of the integer \(n\).

That single fact changes the game. With the scale fixed, predicting a mass is choosing one integer from a fixed ladder of allowed values. There is no continuous dial to slide the prediction onto the measurement. Either an integer lands near the measured mass or it does not.

And the ladder is widely spaced. The simplex numbers grow fast — at the muon's level in the \(d=6\) sector, neighbouring rungs differ by 17%; at the tau's level in \(d=10\), by 43%. So picking the nearest integer does not guarantee a close hit. If the measured value fell at a random spot between two rungs, you would expect the nearest rung to miss by something like a quarter of the gap — several percent, not several parts per million.

This gives a clean way to score each prediction. Take the fractional error \(\varepsilon\) between IDWT and experiment, and compare it to the grid spacing \(g\) at that rung. The ratio \(2\varepsilon/g\) is, roughly, the chance that the nearest available integer would land that close to a randomly placed target. Small means lucky — a measured value sitting almost exactly on a rung, which a flexible fit has no reason to produce.

To keep this honest, restrict to quantities that are both precisely measured and free of any continuous parameter: dimensionless mass ratios, where the sector scale cancels entirely. Each is fixed by a single integer. Treat every mode index as if it were a free choice — the most skeptical possible stance, granting the theory no help from its claim that the indices are forced.

The tau row uses the current PDG measurement (\(1776.93\pm0.09\) MeV), and its \(\varepsilon\) is the measurement's own error bar — the prediction lands inside it, on a grid rung spaced 43% from its neighbours. Multiplying the four independent luck factors gives a joint figure of about \(1.1\times10^{-10}\) — roughly one in nine billion. The lepton ratios carry most of the weight: a hit to better than one part in \(10^4\) on a grid spaced by tens of percent is the part that a loose fit cannot manufacture.

The IDWT predictions span from sub-meV neutrinos to the 172.6 GeV top quark — twelve orders of magnitude from a single formula \(m(n,d)=S(n,d)\times m_{\text{scale},d}\). The chart below shows each predicted mass as a solid bar (log scale); the tick mark at the right end of each bar is the PDG measured value. At this scale the two are visually indistinguishable — the agreement is at the sub-percent level for every particle.

Multiplying luck factors is an estimate, not a likelihood, so the claim is put through two proper statistical tests (computed in idwt.py STEP 39). The first makes the "luck" idea exact. If the measured values had nothing to do with the simplex grid, each would sit at a random position between two rungs — that is the null hypothesis, and the probability of the whole set sitting as close to the rungs as it does can be computed exactly. The quantity list is also widened by a rule fixed in advance: every dimensionless, parameter-free quantity whose measurement is precise enough to tell neighbouring rungs apart. That brings in the Cabibbo angle, the strange-to-down quark mass ratio, the top-to-charm ratio, and the neutrino mass-splitting ratio, for eight quantities in all. The verdict: a joint probability of \(9\times10^{-11}\) — six sigma — or \(1.5\times10^{-7}\) (five sigma) under the most conservative treatment of the measurement errors. Allowing generously for a hundred quantities that might have been examined instead still leaves about \(10^{-8}\).

The second test is the direct one: build random rival theories and let them compete. Each rival keeps everything IDWT derives — the sectors, the scales, the coupling structure — but has its mode indices drawn at random from all values that keep the particles below 1 TeV (neutrinos below 1 eV). Every quantity is then recomputed for the rival, including the three neutrino-mixing angles, and the rival's overall agreement with experiment is scored against IDWT's. In 8.4 million random theories, none came close: the best fell short by a factor of about \(10^{15}\) in joint residual, and the result is unchanged when the index windows are halved or doubled. The tau-to-electron ratio alone was never matched in any draw.

It does not prove IDWT. The result is conditional on the scales genuinely being derived rather than quietly refit — which they are, in the published computation. The clean quantities were chosen by a principled rule (precisely measured, parameter-free), not by closeness; the theory also has looser predictions, and those stand alongside these rather than being hidden. And the set contains real tension as well as agreement: the strange-to-down ratio sits 1.5 sigma from the lattice value, and it is scored as found.

What it does show is enough to retire the first objection. The spectrum is not a "fit anything" structure. Even handed every index for free, the precisely-measured dimensionless ratios sit on the simplex grid far too tightly for a flexible formula — and random integer assignments from the same family never reproduce the data at all.

There is a sharper conclusion underneath the number, and it points at the one thing still open. The whole estimate above bent over backwards to treat each mode index as a free choice. But IDWT's actual claim is stronger: the indices are not chosen at all — they are forced from the three integer seeds \(\{n_d{=}1, n_u{=}3, n_{\rm top}{=}72\}\) (composite \(n_s=4\)) by the generation-tower arithmetic. If that forcing holds, the accounting collapses to three integer seeds and a mass unit producing the entire spectrum, and the significance is no longer one in billions but essentially total. If it does not, the masses are a constrained fit and only the tight grid hits carry weight — still the six-sigma result above, but no more.

So the entire evidential weight of the framework funnels into a single question: are the mode indices forced, or fitted? That is exactly the part IDWT marks as not yet fully derived. This analysis does not answer it — but it shows why it is the question worth everything. Settle the index-forcing and an already-improbable postdiction becomes a theory of the spectrum with almost nothing put in by hand. Everything else is commentary next to that.

The Generation Tower — how the mode indices are built from the two seeds, the forcing this analysis hinges on.

The Simplex Number — what \(S(n,d)\) counts and why its ladder grows fast.

Why \(n_s = 4\) — the composite integer the scales and indices descend from, and the seeds \(\{n_d=1, n_u=3\}\) that generate it.

The Particle Map — the full spectrum these ratios are drawn from.

Mass & Neutrino Predictions — the measured quantities, with errors.