What You’re Testing
The central claim of Astromatics is that celestial angular geometry corresponds to the topological type of human collective events. The engine makes this classification from geometry alone — it does not know what the event is. You provide the date. You provide the event. Compare the output to what actually happened.
- ◈Engine classifies 25 pre-specified topology types from planetary positions
- ◈Zero trainable parameters. No neural network. No learning from your input.
- ◈Ephemeris: Swiss Ephemeris real geocentric tropical positions. No synthetic data.
- ◈Past dates only. Future classification is a paid product.
Run a Test
Enter a past date. The engine classifies geometry only.
Must be before today.
Not sent to the engine — for your comparison only.
Computing planetary positions · Running E₈ classification · φ-coherence scoring...
Engine Output — Geometric Classification
Key Papers — Direct Links
Paper XII · E8 Resonant Systems
Three Physical Substrates · p = 10⁻¹²
Solar system, Saturn moons, Earth tides produce E₈-aligned eigenvectors at 2.4× random. Reproducible with public data.
Paper VII · Topology Proven
202-Event Blind Test · 100% Topology Recall
Blind pre-registration. 202 events, 100% topology recall. Full corpus: 594 events, 4 corpora, zero misses.
Paper XX · Event Sharpness
ρ = 0.770 · p = 0.009 · What ACE Detects
ACE signal correlates with information concentration at event date. Hard-onset events produce strongest signal.
Lean4 Verification
ARAMVerification.lean · 11 Theorems · Zero Sorry
Machine-verified. Aspect-E₈ Identity confirmed: all 5 classical aspect angles are algebraically identical to E₈ inner product classes.
Paper XVII · Correction Log + Lean
Methodology Transparency · Parity Extension
Full correction log, parity extension analysis, and Lean4 verification disposition.
All Research
Complete Research Record · osf.io/ym4tf
All 24 working papers, validation corpora, and supplementary materials.
Reproduce the Core Computation
The Layer 0 result requires only public orbital data and standard linear algebra. Any physicist with Python can reproduce 86.1% E₈ eigenvector alignment vs 35.4% random — 2.4× excess from orbital mechanics alone.
# Paper XII — E8 alignment from orbital periods. Source: osf.io/ym4tf/files/p7gjn
import numpy as np; from scipy.linalg import eigh
periods = {'Mercury':87.969, 'Venus':224.701, 'Earth':365.256,
'Mars':686.971, 'Jupiter':4332.59, 'Saturn':10759.22,
'Uranus':30688.5, 'Neptune':60182.0}
def build_rsm(p, N=15, e=0.03):
b=list(p.values()); n=len(b); H=np.zeros((n,n))
for i in range(n):
for j in range(n):
if i==j: continue
for p2 in range(1,N+1):
for q in range(1,N+1):
d=abs(b[i]/b[j]-p2/q)/(p2/q)
if d<e: H[i,j]=max(H[i,j],1-d/e)
return H
vals,vecs=eigh(build_rsm(periods)); dom=vecs[:,-1]
r=np.ones(8)/np.sqrt(8); print(f"E8 alignment: {abs(np.dot(dom,r)):.4f}")
# Expected: ~0.944