Answers are grounded in this paper and cite their sources.
Description A quantity has one true value, and no instrument reads it: each reads it through a known map, with an offset its maker knows at best, and with noise. This paper is a guide to how the true value is recovered from such readings, written for the numbers the Crystal Topos series is compared with. It takes its name from the gunsmith scene of The Man with the Golden Gun, in which a rifle balanced for three fingers lands an inch low in a five-fingered hand and nothing is wrong with the rifle or the aim. The condition. With every offset free, no number of readings recovers the truth. Three devices restore it: a limit in which the offset vanishes, independent instruments whose offsets scatter, and a closure relation the readings must satisfy together. A reading in a convention has no offset and needs a conversion, which has its own error. The sign. The bodies that publish values do not print the same thing under the plus-or-minus sign. The Particle Data Group's light-quark errors are at 90 percent, a factor 1.645 wider than one standard deviation. The recommended error of the fine-structure constant is an expansion of its inputs by 2.5; the same three determinations give an error of 8, 21, 33 or 37 in the last digits according to the combination rule. Scored against later measurements (Bailey 2017), a stated five standard deviations between two measurements of one quantity has been worth two to three. The instruments. Line-by-line error budgets from the primary sources: a four-flavour lattice calculation of the strange-quark mass (dominant line: the strong coupling, 0.56 of 0.74 percent), the caesium and rubidium atom-recoil determinations of the fine-structure constant (dominant line in both: the phase of a focused beam), the electron's magnetic moment, the muon lifetime, the electron, muon and tau masses, the Z and W masses and the top. Every instrument applies a correction 3 to 57 times its quoted error. Two worked disagreements. The rubidium and caesium values differ by 5.5 standard deviations; a ratio in which every shared input cancels places the gap in a measurement of h/m, and the test that would decide it, the other atom in either apparatus, has not been published. The electroweak vacuum value has six definitions spanning 3 percent; solved as the one common scale of seven physical masses, the series' mass formula asks for 246.56(11) GeV while the series' own derivation gives 245.17 GeV, a disagreement of 0.56 percent that is internal to the series and is not a convention. The map. Where the instrument is coarse or its quantity is a convention, the series' residuals lie within its reach. For the electron, the muon and the Z they do not. The paper claims nothing new about the tower. Receipts: bond_receipts.py, 65 checks, Python standard library only; Bond.lean and Bond.agda check the identifiability claims and 107 integer relations behind the printed figures in the Lean 4 and Agda kernels. I am targetig December 1, 2026 for the first explainer video in 3brown1blue format using manim. You can follow here to get notified when the first video drops https://www.youtube.com/@CrystalToe NOTE: The purpose of this paper is to ground an LLM for The Masses paper. https://doi.org/10.5281/zenodo.23541978 I also highly advise to get the transcripts below for imporved grounding. Now to further ground a LLM I actually get the transcript from Introduction to Lattice QCD (Michael Creutz): Lecture I. Click on the description and you will see "Show transcript." Do the same for Theoretical physics: insider's tricks from Don Lincoln. Upload to your LLM. The mathematics of this paper sits on top of the six foundational Crystal Topos papers. Load them into the model before this paper, as separate uploads in separate turns: a single bundle is compressed into summaries and the algebraic detail is lost. Confirm each — "please confirm you have read this paper and can state its main result" — before sending the next. Load these six first, in this order, one upload per turn: Discrete Core — the algebra , the seed , and its dual. https://doi.org/10.5281/zenodo.20232800 MERA Scaling Spectrum — the two rescalings, the eigenvalues and the multiplicities . https://doi.org/10.5281/zenodo.20236511 Type II Core — the von Neumann factor type and the semi-finite trace. https://doi.org/10.5281/zenodo.20257603 Triangular Duality — the three structures as one categorical object. https://doi.org/10.5281/zenodo.20257953 The Eight Towers and the Foliated Grid — the grid, the eightfold cluster and the 43-layer tower from the spectral data. https://doi.org/10.5281/zenodo.21475952 The Spacetime Constants , , from Substrate Adjacency — the units; the two speeds of the circuit. https://doi.org/10.5281/zenodo.20485003 Then, for the vacuum, the anchor, the lattice and the trees: The Higgs Boson as the Singlet Projector — the vacuum. https://doi.org/10.5281/zenodo.20422007 The Cube Mass Generator — the Yukawa anchor. https://doi.org/10.5281/zenodo.21470949 The Design Tower — the lattice. https://doi.org/10.5281/zenodo.22079045 Pronic Growth — the lattice. . https://doi.org/10.5281/zenodo.19909370 Then this paper. Questions about why the rings, the signs, the floors and the one column tier take the form they do will then be answered against the construction rather than by pattern-matching. Nlte Copyright © 2026 Daland Montgomery. All Rights Reserved. Creative Commons Attribution Non Commercial No Derivatives 4.0 International For a private, proprietary license, contact: quidbit@icloud.com Software Implementation: The formulas and constants derived in this work are implemented in the CrystalAgent engine, available under the AGPL-3.0 Keywords measurement uncertainty, systematic error, fine-structure constant, atom interferometry, Gouy phase, lattice QCD, quark masses, renormalisation scheme, electroweak vacuum expectation value, Fermi constant, lepton masses, W boson mass, CODATA, Particle Data Group, FLAG, Birge ratio, scale factor, expansion factor, identifiability, Crystal Topos
No takes yet. Share an insight, caveat, or question.
Daland Montgomery (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: