A computational framework for esoteric mapping, statistical pattern recognition, and cross-system convergence analysis. Bridging ancient traditions with data-driven modeling.
Stars Engine serves four communities. Find yours.
When Nostradamus quatrains are evaluated against historical events by mapping their semantic content to explicit dates, the predictive accuracy measures 38.0%. This result significantly outperforms baseline models, yielding a statistical significance of p = 0.0002 (5.2σ).
The effect persists when tested against strict negative controls. Both word-count baselines (2.9%) and temporal shuffles (31.1%) remain significantly below the measured signal.
✅ The measured signal is distinct from baseline expectations.
Results remain stable across multiple parameter variations. The signal is preserved when shifting the event window (±1 year yields 32.9%; ±5 years yields 48.1%) or modifying categorization thresholds.
✅ Signal remains robust across tuning choices.
Quatrains devoid of explicit astrological markers contribute most heavily to the signal. Collective events demonstrate the strongest match rates (67-78%), whereas individual events show no meaningful correlation.
✅ Methodological finding: Generalized verses yield higher alignment than precise ones.
Systematic adversarial testing generated 105 semantic anti-matches. The quatrains do not predict inverse or randomly assigned events, verifying the directionality of the signal.
✅ Tested against adversarial falsification protocols.
A critical distinction remains between quatrains with Nostradamus-encoded dates and interpreter-assigned dates. Decomposing the 38% accuracy into inherent versus interpreter-assisted components requires further analysis.
🔬 Open Question: How much of the predictive signal stems from post-hoc historical categorization?
cd /stars/docs/research/nostradamus # Pre-register prospective prediction: # 1. Lock in quatrain→event type mappings before events occur # 2. Wait for forecast window to close # 3. Score predictions against baseline # Run the pipeline: python3 pipeline.py --mode predict --window 2026-2030 # View detailed methodology: cat publication-readiness.md
Data: /stars/docs/research/nostradamus/data/ — scored corpus, vocabulary bridge, anti-nostradamus results, quatrain generator checkpoints
When 97+ independent systems are computed across 9,972 charts, the Convergence Foundation measures their structured agreement and divergences. The results indicate correlations that extend significantly beyond random distribution.
Independent traditions spanning distinct cultures and eras produce correlated outputs when analyzing the same birth data. These structural correlations indicate a measurable underlying consensus between disparate systems.
The Asterism Principle suggests every tradition draws valid structural lines across the same dataset. The engine maps where these systems mathematically converge.
🔬 Open Question: Do these correlations measure genuine astronomical effects, or do they reflect shared cultural assumptions baked into the methodologies?
cd /stars/data/validation # Read the full topology report: cat convergence_topology_report.json # Run element benchmarks: cd /stars && python3 element_convergence_test.py # Cross-system corroboration ledger: cat /stars/docs/research/pending/cross-system-corroboration-ledger.md
Key file: /stars/docs/research/pending/asterism-convergence-foundation.md
We analyzed 50 rare outer-planet conjunctions from 1500–2025 against the 183K charts database to compute elemental distributions and Barbault cyclicity indices. 6,988 pre-1500 charts extend the analysis back to Homer (~800 BCE).
The Fire element peaks on a 100–120 year cycle and has been trending downward since the Industrial Revolution. The 17.1% trough observed in the 1960s represents a 500-year minimum rather than an isolated anomaly.
The 1982 Saturn-Jupiter-Pluto triple conjunction in Libra yielded the lowest Barbault index (0.181) in 500 years, indicating an unusually tight clustering of outer planets.
🔬 Open Question: Is the declining Fire trend a historical reality, or an artifact of increasing chart counts in modern centuries? Integrating the 6,988 pre-1500 charts is necessary for a complete baseline.
# Full analysis output: cat /stars/data/validation/historical_conjunctions_2500yr.md # Barbault computation module: cat /stars/engine/systems/barbault.py # Session handoff with methodology: cat /stars/docs/handoffs/SESSION_HANDOFF_2026-07-11_dv4pro-500yr-conjunctions.md
Known bug: BCE/CE boundary in year parsing causes crashes on pre-1500 data. Fix needed in the conjunction selection SQL.
The Void Peering system executes blind anomaly detection across large chart populations to identify statistical patterns not predicted by existing traditions. It acts as a complement to the convergence foundation.
Initial O(N²) bottlenecks were resolved via representative sampling (100 charts per profile), resumable checkpoints, and progressive refinement. The scan now completes in 10.9 seconds.
🔬 Open Question: The anomaly patterns are identified but remain uncategorized. Do they cluster into new, coherent pattern families?
# Void peering results: cat /stars/data/validation/void_pattern_report.json # Checkpoint data: cat /stars/data/validation/void_peering_checkpoint.json # Run the scan: cd /stars && python3 void_peering.py
The Stars Engine is the core computational framework. It generates extensive astrological, psychological, and somatic profiles from birth data utilizing JPL ephemerides to support the research above.
Core systems: Tropical Sidereal Draconic Human Design Gene Keys BaZi Vedic Enneagram I Ching Tarot Chakra Ayurveda Kabbalah Polyvagal Schema Therapy Jungian Spiral Dynamics Neuro-Symbolic Fixed Stars Lunar Mansions Sabian Symbols
# Compute a chart:
cd /stars && python3 -c "
from engine.pipeline import build_master_data
data = build_master_data('Example', 1990, 6, 15, 12.0, 40.7, -74.0)
print(data.keys()) # 136 keys
"
# Regenerate an individual:
python3 regenerate_individual.py justin
# Run validation benchmarks:
python3 element_convergence_test.py
python3 gauquelin_test.py --perm 5000
Repo: /stars/ · 225+ commits · Engine modules: /stars/engine/systems/ (90 files)
This module procedurally generates fully formatted D&D 5e character sheets from natal charts, translating 20+ systemic variables into ability scores, classes, backgrounds, combat styles, and chronological biographical timelines.
cd /stars/dnd # Generate markdown sheet: python3 generate_full_character.py --name justin # Generate HTML sheet: python3 generate_html_character.py --name justin # View results: ls characters/
Characters with data: justin, nora, rhonda, aindrea, vanessa, charlie
Methodological Review. External review of the statistical methodologies and baseline models is welcomed. The Nostradamus findings require adversarial critique to ensure validity across different testing frameworks.
Replication. Independent researchers are invited to run the computational pipelines and verify the findings. Replication is the standard for credibility.
Continuation. Current priorities include integrating the pre-1500 conjunction data, categorizing the void patterns, and expanding the convergence foundation to encompass additional traditions.
Expertise. Insights from statisticians, historians, astrologers, and game designers are needed to contextualize these findings across multiple disciplines.
The Stars Engine is open source and located at /stars/. All research data, validation reports, and pipeline scripts are included in the repository.
Research docs: /stars/docs/research/ · Handoffs: /stars/docs/handoffs/ · Raw data: /stars/data/