Ancient Greek Astronomical Computer From Antikythera Shipwreck Reveals Sophisticated Mechanism; a 2025 Simulation Study Raises Questions About Whether It Could Run
The Antikythera mechanism, a 2,100-year-old Greek astronomical calculator with at least 39 gears that modeled lunar cycles and predicted eclipses, was recovered from a Mediterranean shipwreck in 1900, though a 2025 simulation study claims manufacturing imperfections would cause jamming in 90% of tests.

- 1The mechanism was discovered in 1900 by sponge divers and identified as a mechanical device in 1902 when museum director Valerios Stais noticed gear-like teeth in the corroded bronze.
- 22005 CT scanning revealed at least 39 gears forming a lunar-solar calendar with a pin-and-slot mechanism modeling the Moon's elliptical orbit and predicting eclipses using the 223-month Saros cycle.
- 3Derek de Solla Price's influential 1974 analysis contained tooth count errors of up to 25 percent, and the highest-resolution 2005 CT images had errors corrected in 2018.
- 4A 2025 simulation study found the mechanism would jam in 90% of tests due to manufacturing imperfections, though this finding has not been corroborated by established research institutions.
- 5The mechanism's maker is unknown, with Hipparchus and Archimedes proposed as candidates, and construction is estimated between 150-60 BCE, likely on Rhodes.
The Full Investigation
5 sections · 5 min read
Construction Date and Origin
According to Communications of the ACM, researcher Michael Edmunds estimates the mechanism's construction date as 140-70 BCE as the preferred period, with a broader range of 150-60 BCE. Multiple Wikipedia sources report the mechanism was constructed approximately 150-100 BCE, with one dating it to about 2200 years ago. These reported ranges overlap substantially and are broadly consistent. Academia.edu reports the mechanism was associated with the School of Poseidonius of Rhodes and dated to circa 87 BCE — a more precise outlier whose basis is not explained by the source.
According to Communications of the ACM, the mechanism was probably built on Rhodes near the Greek philosopher Poseidonius. A University of Zagreb Wikipedia mirror reports that scholarly consensus holds the mechanism was made in Greece with all instructions written in Greek. Academia.edu reports the mechanism was associated with the School of Poseidonius of Rhodes and dated to circa 87 BCE.
Regarding the maker's identity, the World History Encyclopedia reports that the Antikythera Mechanism Research Project attributes the device to Hipparchus, citing his invention and improvement of astronomical instruments and his contemporaneous lifetime. The same source reports that Archimedes is the other popular choice as the mechanism's inventor. No material evidence directly links the device to any specific individual.
Open: The precise date of construction—estimates range from 150-60 BCE with varying preferred windows, and one source reports 87 BCE, but no definitive archaeological evidence settles the question
The Shipwreck
Woods Hole Oceanographic Institution reports that the shipwreck dates from 70-60 BCE and was thought to be carrying luxury cargo from the Asia Minor coast west to Rome. YouTube sources report the same 70-60 BCE dating based on coin and amphora analysis. Greek Wikipedia reports a shipwreck window of 76-60 BCE, potentially up to 50 years after the mechanism's construction. Αρχαιολογία Online reports the ship sank around 50 BCE.
Woods Hole Oceanographic Institution reports that the ship was up to 50 meters long, making it the largest ancient shipwreck ever discovered. YouTube sources report that an earlier theory attributing the cargo to war loot has been replaced by the more probable explanation of a commercial voyage with cargo from diverse Eastern Mediterranean suppliers. A University of Zagreb Wikipedia mirror reports the hypothesis that the mechanism was aboard a ship traveling from Rhodes to Rome.
Wikipedia reports that hull planks were made of elm wood and were carbon dated in 1964 to 220 BCE ± 43 years. The same source explains the discrepancy between this 1964 carbon dating and the 70-60 BCE coin and ceramic dating through the old-tree effect, where the sample came from an old tree cut earlier than the ship's sinking. The Swiss School of Archaeology in Greece reports that its 2025 excavation recovered hull fragments of elm and oak preliminarily dated to around 235 BCE. The 2025 preliminary date of ~235 BCE falls within the 1964 measurement's error range (177-263 BCE), meaning the two datings are consistent rather than contradictory. The institution reports that these 2025 planks are thinner than those found by Cousteau in 1976, raising questions about whether they are from the upper hull section, a repair, or a smaller accompanying vessel.
Open: Why the 1964 carbon date of 220 BCE ± 43 years and the 2025 preliminary date of around 235 BCE both predate the shipwreck by over a century, and whether the old-tree effect fully accounts for this discrepancy or whether other factors are involved
Discovery and Initial Identification
SISSA Conference Proceedings reports that the discovery occurred on Holy Tuesday 1900 (Julian calendar) by Symiot sponge divers. The same source reports that Elias Lykopandis (also called Stadiatis) made the initial discovery at a depth of 35 fathoms and brought up an oversized right forearm from a bronze statue as evidence. ABC News Australia reports that Elias Stadiatis discovered the wreck in 1900 at about 50 meters depth and initially mistook sculptures for rotting corpses and horses. ABC News Australia reports that Captain Dimitrios Kontos verified the find by retrieving a bronze statue arm.
The National Archaeological Museum Athens reports that the first major underwater archaeological expedition was undertaken in 1900-1901 by sponge divers with the assistance of the Greek Royal Navy. The museum reports that Jacques-Yves Cousteau's Calypso conducted the second research expedition to the Antikythera wreck in 1976.
Grokipedia reports that on May 17, 1902, Valerios Stais, while cleaning artifacts at the National Archaeological Museum in Athens, noticed gear-like teeth in a corroded bronze slab and recognized it as a mechanical device, marking the first official identification of the Antikythera mechanism. Wikipedia reports that in 1902, during a visit to the museum, Greek politician Spyridon Stais noticed the artifact containing a gear, prompting his cousin Valerios Stais, the museum director, to conduct the first study. Sources differ on the precise details of the 1902 identification — one attributes the discovery of the gear directly to Valerios Stais, another to his cousin Spyridon Stais prompting Valerios — but both agree the identification occurred in 1902 via the Stais family. The May 17 date rests on the weaker Grokipedia source.
Scientific Analysis and Evolving Understanding
The American Philosophical Society reports that Derek de Solla Price published 'Gears from the Greeks' in 1974, which included collaborative analysis with Karakalos radiographs. The same source reports that all of Price's tooth counts were incorrect, sometimes by as much as 25 percent.
UCL Discovery reports that 2005 imaging studies revealed previously illegible gears and inscriptions, but that the highest-resolution 2005 CT images had some errors. The same institution reports that 2018 reanalysis of the original 2005 X-ray data produced improved image quality, contrast, and legibility.
SISSA Conference Proceedings reports that in early 2005, the Antikythera Mechanism Research Project led by Tony Freeth and Mike Edmunds obtained permission to CT scan the fragments, revealing a complex lunar-solar calendar with a pin-and-slot mechanism modeling the Moon's elliptical orbit and predicting eclipses using the 223-month Saros cycle. Academia.edu reports that the mechanism contained at least 39 gears with Greek inscriptions and displayed scientific scales.
ABC News US reports that a 2008 study revealed a dial predicting Olympic games timing (first new moon after summer solstice) every four years, which historian Alexander Jones of NYU called an unexpected connection between astronomy and mundane games.
Open: Whether the 2025 Szigety-Arenas simulation accurately models ancient operating conditions, including manufacturing tolerances, lubrication, materials, or operating force. The finding is a single uncorroborated simulation reported by one outlet, and the absence of corroboration or endorsement from established Antikythera research groups is itself a relevant data point about its current standing
A 2025 Simulation Study
Artnet News reports that a 2025 study by Argentina researchers Szigety and Arenas found that the mechanism would jam or desynchronize in 90% of simulation tests due to manufacturing imperfections with triangular teeth that caused gears to lose contact, typically stopping before completing four months of solar motion. This finding has not been corroborated by other sources or peer-reviewed research institutions involved in Antikythera mechanism analysis. The available source does not specify whether the simulation accounted for materials, lubrication, operating force, or actual ancient manufacturing tolerances.
Open: Whether the 2025 Szigety-Arenas simulation accurately models ancient operating conditions, including manufacturing tolerances, lubrication, materials, or operating force. The finding is a single uncorroborated simulation reported by one outlet, and the absence of corroboration or endorsement from established Antikythera research groups is itself a relevant data point about its current standing