Speed-Error Cross-Correlation Dating of Ancient Star Catalogues, with Application to the Almagest
Carlos Baiget Orts · 2026
Read Speed-Error Cross-Correlation Dating of Ancient Star Catalogues, with Application to the Almagest by Carlos Baiget Orts (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-02521.txt.
Introduction
Baiget Orts presents SESCC, which dates ancient star catalogues through the link between stellar speeds and position errors. At the true date the two are unrelated, so the best date is the trial date that minimizes their correlation. Validated against the catalogues of Tycho Brahe and Ulugh Beg, the method applied to the Almagest points mostly to pre-Christian dates, favoring a Hipparchan origin over a Ptolemaic one.
Catalog record
| CREATOR | Carlos Baiget Orts |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26019 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.02521
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Speed-Error Cross-Correlation Dating of Ancient Star Catalogues, with Application to the Almagest Carlos Baiget Orts Abstract We present SESCC (Speed-Error Signals Cross-Correlation), a method for dating ancient star catalogues from the cross-correlation between stellar proper-motion speeds and positional residuals. At the true epoch, residuals are independent of proper-motion speed; the epoch estimate is the trial date that minimises this cross-correlation. For ecliptic latitudes, SESCC applies the dot product between speeds and residuals across all catalogue stars without subset selection or linear modelling. For ecliptic longitudes, SESCC-pairs uses pairwise longitude differences between neighbouring stars, making the method immune to any global longitude offset by algebraic construction. Validated against Tycho Brahe (1547 CE, true ∼ {\sim} 1580 CE) and Ulugh Beg (1452 CE, true 1437 CE), and confirmed invariant under offsets of ± 6 ∘ \pm 6^{\circ} , the method is applied to the Almagest. Both coordinates yield bootstrap distributions with 74% pre-Christian minima, consistent with a Hipparchan origin and inconsistent with a Ptolemaic one. The near-absence of quarter-degree fractions in the Almagest longitudes, explained as the deterministic consequence of Ptolemy’s precession correction, provides independent corroboration. Keywords: Almagest; Hipparchus; Ptolemy; star catalogues; proper motion; epoch dating; precession; sexagesimal fractions 1 Introduction The question of when an ancient star catalogue was observed has occupied historians of astronomy for over three centuries. The ecliptic positions of stars change over time due to proper motion, the slow individual displacement of each star across the celestial sphere. A catalogue compiled at epoch T_{0} will show residuals — differences between catalogued and modern positions — that are minimised when the trial epoch T matches T_{0} . This is the basis of all proper-motion dating methods. For most historical star catalogues, the epoch of observation is known from independent sources with sufficient precision for verification purposes. The catalogue of Tycho Brahe reflects observations between 1576 and 1597 CE ( 5 ) . The catalogue of Ulugh Beg was compiled at his observatory in Samarkand, with the epoch established as 1437 CE from manuscript sources ( 10 ; 19 ) . These catalogues therefore serve as natural validation benchmarks for any dating method: a method that fails to recover their known epochs cannot be trusted when applied to catalogues of uncertain date. The Almagest star catalogue presents a different situation. Ptolemy assigns it an epoch of 137 CE, the reign of Antoninus Pius, but the catalogue’s systematic errors have led many investigators to question whether the positions are genuinely Ptolemaic or derive instead from earlier observations by Hipparchus of Nicaea, whose activity is documented between 162 and 127 BCE. The longitude errors show a systematic offset of approximately -1^{\circ} relative to the Hipparchan epoch, consistent with Hipparchan positions adjusted to 137 CE using Ptolemy’s erroneous precession constant of 1^{\circ} per century rather than the correct value of approximately 1.4^{\circ} per century. This observation, noted since the time of Tycho Brahe and reviewed extensively by 8 , motivates the application of proper-motion methods to determine an independent epoch estimate for the catalogue coordinates. The earliest systematic proper-motion dating of the Almagest was proposed by 3 , whose ‘bulk method’ applies linear regression between position residuals and proper-motion speed to a selected group of fast-moving stars, yielding T=-89\pm 112 years. 6 re-examined the bulk method and demonstrated that the underlying linear correlation is not statistically significant in the Almagest data. 13 further showed that the method fails to recover the known epoch of Ulugh Beg, undermining its general validity. 2 used the declinations in Book VII of the Almagest rather than the catalogue proper, recovering an epoch near -128 BCE for observations attributed to Hipparchus. All these methods address latitudes or declinations; longitude-based dating has relied on precession arguments and offset analyses ( 4 ; 11 ; 12 ) that necessarily incorporate assumptions about the global longitude zero-point. We present SESCC (Speed-Error Signals Cross-Correlation), a new proper-motion dating method that addresses these limitations. At the true epoch, positional residuals carry no systematic dependence on proper-motion speed; the epoch estimate is the trial date that minimises this speed-error cross-correlation. For ecliptic latitudes, SESCC applies the dot product between proper-motion speeds and positional residuals across all catalogue stars simultaneously, without selecting a subset and without fitting a linear model. For ecliptic longitudes, we introduce an extension called SESCC-pairs that uses pairwise longitude differences between neighbouring stars. Because any uniform longitude offset — whether from precession correction, instrumental zero-point error, or any other global shift — cancels exactly in the difference between any two stars, SESCC-pairs is immune to global longitude offsets by construction. This makes it the first longitude-based dating method that requires no assumption about precession. The paper is organised as follows. Section 2 presents the SESCC method for latitudes. Section 3 introduces SESCC-pairs for longitudes and establishes its invariance to global offsets. Section 4 validates both methods against the catalogues of Tycho Brahe and Ulugh Beg, and against synthetic catalogues with realistic positional noise. Section 5 applies both methods to the Almagest and discusses the results alongside the fractional position evidence. Section 6 concludes. 2 The SESCC Method for Ecliptic Latitudes 2.1 Formulation Let a star catalogue contain N entries, each identified with a star in the Hipparcos catalogue ( 7 ) . For each star i , let \beta_{i} be the ecliptic latitude recorded in the catalogue, and let \beta_{i}^{\rm mod}(T) be the ecliptic latitude computed from the Hipparcos position and proper motion propagated to trial epoch T . The residual of star i at epoch T is \delta\beta_{i}(T)=\beta_{i}-\beta_{i}^{\rm mod}(T) . The SESCC statistic is: C(T)=\sum_{i}|\mu_{\beta i}|\cdot|\delta\beta_{i}(T)| (1) where |\mu_{\beta i}| is the absolute latitudinal proper motion of star i in milli-degrees per millennium. The epoch estimate is \hat{T}=\arg\min_{T}C(T) . The statistical justification is as follows. At the true epoch T_{0} , the residual \delta\beta_{i}(T_{0}) reflects the measurement error of star i in the original catalogue, which is random and independent of |\mu_{\beta i}| . The dot product C(T_{0}) is therefore the sum of independent random terms with no systematic tendency to grow with |\mu_{\beta i}| . At any other trial epoch T\neq T_{0} , the residual contains an additional term proportional to |\mu_{\beta i}|\cdot|T-T_{0}| , so fast-moving stars contribute disproportionately large values to C(T) . The statistic is therefore minimised at the true epoch and increases as T departs from T_{0} . Stars with small |\mu_{\beta i}| contribute little weight to the dot product automatically, without any explicit selection or threshold. This implicit weighting concentrates the signal in the stars that carry the most temporal information, while stars with negligible proper motion contribute only noise and are naturally down-weighted. Precession does not affect ecliptic latitudes, so the method requires no precession correction. 2.2 Practical implementation Modern positions and proper motions are taken from the Hipparcos catalogue ( 7 ) . Star positions at each trial epoch are computed using Skyfield ( 16 ) with the JPL DE441 ephemeris ( 14 ) . Identifications between catalogue entries and Hipparcos stars follow 19 , whose machine-readable versions of the Almagest and Ulugh Beg catalogues provide ecliptic coordinates, positional errors relative to Hipparcos, and identification quality flags. The trial epoch is scanned from 1900 CE to -600 BCE in steps of one