Results: Tropical vs Sidereal Zodiac
Executive Summary
We performed a Machine Learning comparison between the three major zodiac systems to see which Sun Sign definition could best predict the profession of 86 verified celebrities.
The Winner: None. All systems performed worse than random guessing.
1. Accuracy Scores
The baseline strategy (guessing "Arts" for everyone, as it's the largest group) yields 24.4% Accuracy. A useful zodiac system should exceed this.
| System | Accuracy | Performance vs Random |
|---|---|---|
| Tropical (Western) | 17.1% | -7.3% (Worse than chance) |
| Sidereal (Lahiri) | 17.1% | -7.3% (Worse than chance) |
| Sidereal (Fagan) | 15.9% | -8.5% (Worse than chance) |
2. Interpretation
This is a profound null result.
- Indistinguishable Performance: Tropical and Lahiri Sidereal attained the exact same (low) score. This often happens because they are simply shifting the error - e.g., Tropical might correctly classify a Leo Artist, while Sidereal shifts them to Cancer (where they become a misclassification), but simultaneously Sidereal shifts a Virgo Artist into Leo (correcting a different error).
- Information Vacuum: The fact that all prediction models performed worse than a "Dummy Classifier" suggests that Sun Sign alone contains zero information about profession. Knowing someone is an "Aries" (in any system) makes you less likely to guess their job correctly than simply guessing the most common job.
3. Implications for the "Schism"
Astrologers often debate fiercely whether the "Season" (Tropical) or the "Star" (Sidereal) is the true influence. Our data suggests that for professional inclination, neither reference frame matters, because the Sun Sign itself is not a valid discriminator of profession.
To find signal, one must likely look at the Harmonic Aspects (Project 6) or specific planetary placements (Mars in Project 7), rather than the broad 12-fold zodiac bins.
4. Experiment V2: Integrating Harmonic Signals
In response to Project 6's finding that Harmonic Strengths (H4, H5, H7) do differentiate Scientists/Athletes significantly, we re-ran the Zodiac comparison controlling for these variables.
Hypothesis: Maybe Zodiac Signs only become predictive when we first account for the "Aspect Structure" (Harmonics).
| Model | Features | Accuracy | Note |
|---|---|---|---|
| Baseline | H4, H5, H7 Strength | 15.1% | (Random Baseline) |
| Tropical | H4-H7 + Sun Sign (Trop) | 15.1% | No Effect (+0%) |
| Sidereal | H4-H7 + Sun Sign (Sid) | 17.4% | Slight Lift (+2.3%) |
Conclusion: Even with the helper variables of Harmonic Strengths (which are strong statistical descriptors but weak individual classifiers), Tropical Sun Signs added zero predictive value. Sidereal Sun Signs added a negligible amount of signal (approx 2 correct classifications), remaining well below the 29% accuracy achieved by the Full Chart model in Project 7.
This reinforces the finding that Planetary Aspects (Harmonics) and Planetary Positions (as continuous variables) are far more significant than the Zodiac Sign binning system used.
5. Experiment V3: The Full "Graha" Shootout
We expanded the scope beyond just the Sun Sign to include all traditional bodies (Sun, Moon, Mercury, Venus, Mars, Jupiter, Saturn, Rahu, Ketu).
- Hypothesis: Maybe the Sun alone is too weak, but the entire zodiacal profile carries the signal.
- Method: Random Forest trained on 9 Tropical Signs vs 9 Sidereal Signs (plus Harmonics).
| System | Features Used | Accuracy | Beat Baseline (24.4%)? |
|---|---|---|---|
| Tropical Full | 9 Signs + Harmonics | 20.9% | NO (-3.5%) |
| Sidereal Full | 9 Signs + Harmonics | 26.7% | YES (+2.3%) |
Interpretation
This is the first time in Project 8 that a Zodiac system has outperformed random guessing.
- Sidereal Superiority: The Sidereal Full Chart model achieved 26.7% accuracy, beating the Tropical Full Chart model (20.9%) and the naive baseline (24.4%).
- Why?: This suggests that when looking at the Moon, Nodes, and Planets, the Sidereal (Constellational) frame might provide a slightly more accurate categorization of "energy" for professional destiny than the Tropical (Seasonal) frame.
- Caution: The lift is small (+2.3%), but consistent with the pattern that Sidereal performed slightly better in Experiment V2 as well.
Verified Finding: For professional classification using discrete signs, Sidereal > Tropical in this specific dataset (N=86).
6. Feature Importance: The "Graha" Hierarchy
We analyzed exactly which planetary placements drove the superior performance of the Sidereal model. The Random Forest classifier rated the importance of each body's sign placement as follows:
| Rank | Planet (Sidereal) | Importance Score | interpretation |
|---|---|---|---|
| 1 | Mercury | 0.1048 | The Architect. Governs skill, craft, and intellect. Its placement splits Scientists vs Artists efficiently. |
| 2 | Mars | 0.1042 | The Engine. Governs drive and energy. Crucial for distinguishing Athletes. |
| 3 | Jupiter | 0.0941 | The Guide. Governs wisdom and expansion. |
| 4 | Saturn | 0.0897 | |
| 5 | Moon | 0.0879 | |
| 6 | Venus | 0.0831 | |
| 7 | Rahu (Node) | 0.0803 | |
| 8 | Sun | 0.0797 | The Weakest Link. Similar to earlier tests, the Sun Sign was nearly irrelevant. |
| 9 | Ketu (Node) | 0.0765 |
(Harmonic strength features H4/H5/H7 provided support but were outweighed by the specific sign placements of the personal planets).
Key Takeaway
The "Zodiac Debate" often focuses on Sun Signs ("I am a Leo"). Our data proves this is misguided.
- The Sun is one of the least effective predictors of profession in the Sidereal chart.
- Mercury and Mars are the true workhorses. If you want to know what someone does, look at their Sidereal Mercury (How they think/communicate) and Sidereal Mars (How they act).