Experiment A · The Resonance Atlas
Where does the Earth
still ring?
A comparative-resonance instrument. Curated coordinates from cultural, geophysical, and modern-engineering sites. Pick any classes, any origin, and let a logarithmic-spiral regression refit live. Read the gap between signal and null.
EXPLAIN LIKE I'M 10
· everyday-language one-paragraph version
A world map of special places — twelve old Indian temples, fifteen Egyptian pyramids, twelve Mesoamerican pyramids, the African Cradle, sacred lakes, paleo caves, glaciers, impact craters, waterfalls, salt flats, tectonic triple junctions, hot springs, deep-Earth hotspots, and a control set of modern engineering wonders. Pick which sets to show. Pick where the 'center' is — or let the page find it. A spiral tries to fit the points. The big question is whether real ancient sites fit a hidden spiral better than randomly-built modern buildings would. The page reports the gap as a single number: Δ(signal − null).
WHAT YOU ARE LOOKING AT
A live map and a polar plot of curated coordinates from seventeen classes of site — nine sacred + Swayambhu lineages (Jyotirlinga, Shakti, Egyptian, Mesoamerican, Megalithic, Andean, African Cradle, sacred lakes & rivers, paleo & sacred caves), six natural-anchor classes (hot springs, mantle hotspots, cryospheric anchors, impact craters, waterfalls, salt flats, tectonic triple junctions), and a null set of modern engineering wonders. Pick any class, any origin point, and the page refits a logarithmic spiral to whatever subset is currently active. R² tells you how well the spiral fits each class.
WHY THIS MATTERS FOR QUANTUM
The premise of the simulator is that geometry is the substrate-independent part of the coherence problem. If the same generative geometry (a log-spiral with a particular pitch) shows up across cultures and across natural geophysical features, that is empirical support for a universal generating function — and a starting point for which cavity geometries to test in hardware. If only one class fits the spiral, the claim is local. The number that matters is Δ(signal − null): the gap between sacred / geophysical classes and the modern-engineering control.
WHAT TO WATCH
Toggle classes on and off; switch origin between Narmada, Kailash, Réunion, Giza, or drag your own; press Best-fit to let the data choose the origin. Watch the per-class R² readouts. If modern engineering ever fits as well as Jyotirlinga, the spiral structure is geographic coincidence, not signal.
Active sites
25
2 classes
Combined R² (log)
0.981
r = 2.108 · exp(0.1174·θ)
Origin
21.14°, 76.72°
Narmada centroid
Δ(signal − null)
+0.096
signal R² 0.978 · null R² 0.882
Atlas v3 · generational Fibonacci shell · per-class origin search
One emitter, many shells.
Fk sites per generation, φk radial growth, 137.5° angular step.
Each lineage is fit independently as a Fibonacci shell pattern. Four parameters per fit: (latO, lonO, r0, θ0). We compare each class's recovered origin against the Narmada basin, and emit predicted lat/lon for the missing sites that would complete generation eight (cumulative 88).
Regression sample
v0.34 baseline · the six architectural lineages of the published abstract. Reproduces the Hudson Bay 5.4° / p≈0.025 anomaly-convergence result.
Runs a coarse 5° × refined 0.5° origin search across 6 classes, then 200 null shuffles per class inside each class's bounding box. Around 1.5 – 3 seconds in total.
Geographic projection · equirectangular
Origin: 21.14°, 76.72°
25 sites · 2 classes
combined R² = 0.981
Polar plot · centred at origin
r in arc-degrees
θ = bearing CW from N
θ = bearing CW from N
spiral: r = 2.11 · exp(0.117·θ)
Class fingerprint · each class's natural origin
Where does each class want its centre?
Press "Compute fingerprints" to grid-search the natural origin of every class. Each class runs an independent 2-pass search over the whole world (≈2-3 s per class). If multiple classes share the same natural origin, that's a real signal across cultures.
Two-class delta · R² at every preset origin
Do two classes share an origin?
Class A
Class B
| Origin | Jyotirlinga R² | Modern Engineering R² | Δ (A − B) |
|---|---|---|---|
| Narmada centroid | 0.978 | 0.882 | +0.096 |
| Mount Kailash | 0.940 | 0.873 | +0.067 |
| Réunion hotspot | 0.103 | 0.887 | -0.784 |
| Giza Plateau | 0.018 | 0.880 | -0.861 |
| Mid-Atlantic Ridge | 0.768 | 0.976 | -0.208 |
| Yellowstone hotspot | 0.805 | 0.980 | -0.175 |
Pick two classes. The table is the R² for each at every preset origin. Origins where Δ is large and positive favour class A; large negative favours class B; small Δ means the two classes are competing for similar generative geometry there.
Add custom site
Paste a coordinate · refit lives.
Save & interpret
Snapshot or ask Claude for a reading.
The interpretation reads your current configuration (active classes, origin, per-class R², Δ value) and returns a 4-6 sentence plain-language reading. Same Claude-Sonnet pipeline that powers the other experiment pages.
v3 · null distribution + AIC penalty
Is the observed R² above chance?
Class to test
Trials
The bootstrap shuffles each site within its class's bounding box, then refits with free origin. If the observed R² sits in the top 5% of trial R²s, the spiral structure is unlikely to be a geographic artefact. Hardens the Atlas's headline numbers against the floated-origin-overfits-anything finding.
AIC / BIC · combined fit (4-param origin)
n (sites)25
k (free params)2
SSR (log-space)1.0679
AIC-3.88
BIC-1.44
Lower AIC = better fit per degree of freedom. Floating the origin adds 2 parameters; if R² doesn't improve enough to overcome that penalty, the fixed-origin model is preferred.
Lock the current state into a shareable URL. Anyone clicking it sees the same classes, the same origin, the same numbers. Cite the link in a paper or talk and the screenshot becomes a clickable artefact.