Try Astrologer API

Subscribe to support and grow the project.

Precision Summary #

Compact precision summary for LibEphemeris vs pyswisseph 2.10.03. For full details, models, and methodology see reference/precision.md and reference/swisseph-comparison.md.

Planetary Positions (geocentric ecliptic, 1550–2650 CE) #

Body Mean Diff Max Diff Notes
Sun 0.04" 0.20" DE440 vs DE431
Moon 0.70" 3.32" Numerical vs analytical lunar theory
Mercury 0.05" 0.32"
Venus 0.08" 0.33"
Mars 0.06" 0.58"
Jupiter 0.12" 0.44" Includes COB correction
Saturn 0.13" 0.51" Includes COB correction
Uranus 0.23" 0.50"
Neptune 0.24" 1.17"
Pluto 0.26" 0.75" Includes COB correction

All planets sub-arcsecond. Moon ~3" max reflects different lunar models (JPL DE440 numerical integration vs ELP/MPP02 + DE431).

Velocities #

Component Max Diff
Longitude speed < 0.003°/day
Latitude speed < 0.004°/day
Distance speed < 0.0001 AU/day

Lunar Points #

Point Max Diff Independent Verification
Mean Node < 0.001°
True Node < 0.01" Verified vs JPL Horizons to machine precision
Mean Lilith < 0.015" (lon) Latitude ~20" systematic (different node formulas)
True Lilith < 0.5" Both libraries ~240" from Horizons (inherent two-body limit)
Interpolated Apogee ~0.36° Genuine algorithm difference (JPL DE440 vs ELP2000-82B perturbation series)
Interpolated Perigee ~2.6° JPL DE440 physical passages vs truncated ELP2000-82B perturbation series

House Cusps #

< 0.02" for all 24 supported house systems, tested at 11 global locations. Iterative systems (Placidus, Koch) use 10⁻⁷° convergence threshold.

Fixed Stars #

116 stars from Hipparcos catalog with van Leeuwen 2007 proper motions. Max difference: 0.51" (Rigil Kentaurus — nearest star, parallax not modeled). 98% of 101 comparable stars within 0.5". Two catalog bugs found and fixed (Algedi wrong component, Asellus Borealis wrong HIP number).

Ayanamsha #

  • Standard modes (Lahiri, Fagan-Bradley, Raman): < 0.0002°
  • Star-based modes (True Citra, True Revati): < 0.006°

Eclipses #

  • Solar eclipse timing: < 6 seconds
  • Lunar eclipse timing: < 8 seconds
  • Rise/set/transit: < 30 seconds

Delta T #

  • Modern (1900–2025): < 1 second
  • Historical (< 1700): up to ~187 seconds (different models: SMH 2016 vs E&M 2006)
  • Future (> 2050): grows with extrapolation divergence

Minor Bodies #

  • With SPK kernels: sub-arcsecond (matching JPL Horizons)
  • Keplerian fallback: ~10–30" for main belt asteroids near epoch, degrees over decades

Hypothetical Planets #

Uranian hypothetical planets: < 1" (Keplerian from published elements).

Heliocentric / Barycentric / Equatorial / XYZ #

Mode Max Diff
Heliocentric < 0.0004° (1.1")
Barycentric (non-Sun) < 0.001°
Equatorial RA/Dec < 0.0005° (1.7")
XYZ Cartesian < 0.00005 AU

Hyper-Validation #

4400+ comparison rounds across 29 API sections: 3947 PASS, 441 KNOWN, 0 FAIL, 12 SKIP. All divergences documented in divergences.md.