ORX-LAB-001 · OPEN RESEARCH
If a birth time is off by 30 minutes, how much does the chart actually change?
Birth-Time Uncertainty and Chart Stability Across Bazi, Zi Wei Dou Shu, and Western Astrology: A Computational Sensitivity Study
The short answer
Shift a birth time by 30 minutes and something in the chart changes 97.7% of the time. But what changes is completely different from system to system.
- 2.8%of Bazi hour pillars change. The hour pillar is a two-hour step, so 30 minutes usually survives unless you sit near a boundary.
- 2.8%of Zi Wei ming palaces change — the identical rate, because it is derived from the same two-hour branch boundary.
- 24.2%of Ascendant signs change. The Ascendant slides continuously rather than stepping, so its odds of crossing a boundary are simply different.
- 6.4°is the median shift of the Ascendant even when the sign holds. Every house cusp moves with it.
- 96.5%of charts have at least one planet in a different house. The planets have not moved; the territory they are read in has.
Both ends, for scale: at 5 minutes something changes 43.8% of the time; at 2 hours it is 100.0%.
Why the curves have different shapes
The table below gives the numbers; the chart gives the shape. The shape is the finding. Bazi and Zi Wei sit flat and low until the offset approaches their two-hour step, then jump. Western house placements climb almost immediately, because house cusps are far more densely spaced than a two-hour branch.
- Bazi hour pillar
- Zi Wei ming palace
- Ascendant sign
- Planet houses
| Offset | Anything changed | BaziHour pillar | BaziDay pillar | Zi WeiMing palace | Zi WeiBody palace | AstrologyAscendant sign | AstrologyMidheaven sign | AstrologyPlanet houses | Asc. shift |
|---|---|---|---|---|---|---|---|---|---|
| -120 min | 100.0% | 100.0% | 8.3% | 100.0% | 100.0% | 89.7% | 98.5% | 100.0% | 25.9° |
| -60 min | 100.0% | 52.8% | 0.0% | 52.8% | 52.8% | 52.1% | 45.4% | 100.0% | 12.9° |
| -30 min | 97.5% | 5.6% | 0.0% | 5.6% | 5.6% | 27.6% | 20.8% | 94.7% | 6.4° |
| -15 min | 81.5% | 0.0% | 0.0% | 0.0% | 0.0% | 14.0% | 11.3% | 75.9% | 3.2° |
| -5 min | 43.5% | 0.0% | 0.0% | 0.0% | 0.0% | 5.0% | 3.8% | 39.1% | 1.1° |
| +5 min | 43.8% | 0.0% | 0.0% | 0.0% | 0.0% | 3.5% | 5.0% | 38.5% | 1.1° |
| +15 min | 83.1% | 0.0% | 0.0% | 0.0% | 0.0% | 10.1% | 12.8% | 78.5% | 3.2° |
| +30 min | 97.7% | 2.8% | 2.8% | 2.8% | 2.8% | 24.2% | 26.2% | 96.5% | 6.4° |
| +60 min | 100.0% | 47.2% | 8.3% | 47.2% | 47.2% | 50.9% | 54.4% | 100.0% | 12.7° |
| +120 min | 100.0% | 100.0% | 16.7% | 99.7% | 100.0% | 91.3% | 96.9% | 100.0% | 25.8° |
An unplanned finding
The Bazi hour pillar, the Zi Wei ming palace and the body palace change at exactly the same rate at every offset. That is not three independent systems agreeing. It means two of them are reading the same two-hour boundary.
This matters for anyone designing cross-system validation: treating Bazi and Zi Wei hour-derived values as independent confirmations of each other overstates the evidence. They share an input, so they share a failure mode.
What we measured
- How often each system’s value changes at each offset size
- The offset at which each value first becomes unstable
- How far the Ascendant actually moves, in degrees
- How often at least one planet lands in a different house
What this study does not claim
- We did not measure which system is more accurate.
- This is not a study of predictive power or of whether divination works.
- A stable computation is not the same thing as a correct interpretation.
- The sample is an evenly spread grid of dates and times, not a real birth population.
Method and data
864 synthetic birth times were built from an even grid of 3 cities, 2 years, 12 dates per year and 12 times per day. Each was computed at its true time and again at every offset in -120, -60, -30, -15, -5, 5, 15, 30, 60, 120 minutes, for 9,504 chart computations in total.
All three systems were computed by the same production engines this site uses, so no cross-engine difference can leak into the comparison. The calculation path contains no randomness, no wall-clock time and no external calls: the same command produces the same table.
Citing this study
The dataset is published in a separate repository so that a citation keeps working even if this site changes. That repository carries a CITATION.cff file, so GitHub renders a ready-made citation under “Cite this repository”.