← All discoveries
History / political longevity · 2026-04-13

Sobhuza II's 82-Year Reign Is the Verifiable Wikidata Maximum

Reference works listing the longest-reigning monarch should cite Sobhuza II rather than disputed older entries; the audit reveals which other long-reign claims fail provenance checks.

Description

Queried the public Wikidata SPARQL endpoint at query.wikidata.org for every (person, position) pair where the position is a subclass of Q116 ('monarch') and both the start (P580) and end (P582) qualifiers exist on the P39 statement. The query returned 6,762 reign records, which collapsed into 5,917 distinct (monarch, position) tenures after merging stints under the same title. The full result is pinned by SHA-256 80107e9c1a7937d3ca5718e95bb300de44a4570a16c80436f171d75662282a09. Sorting by tenure length in days reveals two distinct populations at the top of the distribution.

Purpose

Precise

Ledger + a structural data-quality thesis. The ledger is the top-20 longest verified monarch tenures pulled directly from Wikidata. The thesis is that the raw rank-1 through rank-7 entries are all artefacts: a 603-year tenure from Guy de Ventadour as Bishop of Cambrai is a 1952-for-1352 transcription error; a 180-year reign of Sultan Mudzafar Shah I of Kadaram is a legendary record from a 10th-12th century chronicle with no day-precision evidence; a 106-year bishopric, a 101-year unnamed Q133854890 entry, and 97-year, 93-year, 86-year reigns from Cambodian, Korean, and legendary Japanese chronicles all use Jan 1 placeholder dates that would round any year-precision range to a maximal day count. The first day-precision, independently verifiable, modern-data entry is rank 8: Sobhuza II of Eswatini at 30,204 days (82.74 years) from 1899-12-10 to 1982-08-21. This matches the widely-cited 'longest verified reign in human history' figure. After filtering errors, the second-longest is Louis XIV of France at 26,407 days (72.30 years) from 1643-05-14 to 1715-09-01 — leaving a 3,797-day (10.4-year, 14 %) gap between rank 1 and rank 2, the largest single gap anywhere in the top-10. Below Louis XIV the entries cluster tightly within a few months of each other: Johann II of Liechtenstein 71.0 y, Elizabeth II 70.6 y (listed four separate times in Wikidata for her four crowns), Bhumibol of Thailand 70.3 y, Shapur II of Sasanian Persia 70.0 y. The data-quality subobservation is itself part of the result: any naive 'longest reign' analysis using Wikidata SPARQL needs explicit error filtering, otherwise the top-7 results are all wrong.

For a general reader

Who reigned the longest of any verified monarch in history? You'd think this would be a settled question — and it is, but only if you check the answer carefully. I asked Wikidata, the open structured-data sister of Wikipedia, for every monarch with both a start date and an end date for their reign. It returned almost 6,000 entries. When I sorted by reign length, the very top of the list was nonsense: a 603-year-long French bishop, a 180-year sultan, a 100-year-long unnamed monarch, and several legendary kings whose dates were just '01-01' placeholders for 'sometime that century.' Specifically the top SEVEN results were all junk. The first real entry, the first one with actual day-precision dates and a real biography, is Sobhuza II, who was King of Eswatini (formerly Swaziland) from December 10, 1899 to August 21, 1982 — 82 years and 8 months. That matches what reference works have always said: Sobhuza II is THE longest-reigning monarch in human history with reliable records. The second-longest is Louis XIV of France, the 'Sun King,' at 72 years and 3 months, which leaves a 10-year gap between him and Sobhuza II. Ten years is a LOT of margin in this sport — below Louis XIV the next several monarchs (Liechtenstein's Johann II, Elizabeth II of the UK, Bhumibol of Thailand) all clock in within a few months of each other around the 70-year mark. So the headline is: Sobhuza II is way out front; the runners-up are bunched together; and Wikidata's top-7 raw query results are wrong — even something as well-defined as 'longest reign' has data quality issues that you have to clean up before you can trust the answer.

Novelty

The Sobhuza II 'longest reign' fact is famous; the Wikidata data-quality issue with the top-7 raw query results is not, and the specific 10.4-year (14 %) gap between Sobhuza II and Louis XIV with the explicit caveat that it represents the largest gap anywhere in the verified top-10 has not been published as a single pinned claim. The pinning to a specific Wikidata SPARQL response by SHA-256 also makes the data-quality finding independently re-runnable.

How it upholds the rules

1. Not already discovered
Wikipedia and reference works list Sobhuza II as the longest-reigning verified monarch, but neither the specific 10.4-year gap to Louis XIV (which is the largest gap in the top-10) nor the systematic data-quality observation about Wikidata's top-7 entries is documented anywhere I could find on 2026-04-13.
2. Not computer science
History / political longevity. The objects of study are documented monarchic tenures; the program is a SPARQL fetch and a sort with explicit filter rules.
3. Not speculative
Every number is read directly from the pinned Wikidata response. The data-quality classification of the top-7 errors is supported by independent verification: Guy de Ventadour's Wikipedia entry shows the corrected 1352 end date; the Sultan Mudzafar Shah I, Cambodian and Korean kings, and Emperor Nintoku entries all use Jan-1 placeholder dates visible in the raw query JSON.

Verification

(1) Wikidata SPARQL response pinned by SHA-256 80107e9c1a7937d3ca5718e95bb300de44a4570a16c80436f171d75662282a09. Anyone can re-fetch by re-running the same query against the same endpoint. (2) The Sobhuza II date pair (1899-12-10, 1982-08-21) and the Louis XIV date pair (1643-05-14, 1715-09-01) are both directly verifiable against any biographical reference and against their Wikipedia entries. (3) The top-7 errors are flagged independently in the data: Guy de Ventadour 1349..1952 is a 603-year span, which is physically impossible for any individual; Sultan Mudzafar Shah I 956..1136 is a 180-year span over a chronicle-only record; Nintoku is in Wikipedia's 'Legendary Emperors of Japan' list. Each error is independently identifiable. (4) Elizabeth II appears at exactly the same length (70.6 years) four times in the top-20 because Wikidata records her crowns as four separate positions (UK, New Zealand, Canada, Australia); this is itself a verifiable structural fact about the data model.

Sequences

Top 8 longest tenures (raw, including data errors)
603.0 y Guy de Ventadour [error] · 180.0 y Mudzafar Shah I [legendary] · 105.7 y Busso von Alvensleben [error] · 101.0 y Q133854890 [placeholder] · 97.0 y Jyesṭhāryā [placeholder] · 93.0 y Taejo [legendary] · 86.0 y Nintoku [legendary] · 82.7 y Sobhuza II [VERIFIED]
Top 8 verified tenures (errors removed)
82.74 y Sobhuza II of Eswatini · 72.30 y Louis XIV of France · 71.00 y Johann II of Liechtenstein · 70.60 y Elizabeth II (UK/NZ/Canada/Australia, 4 entries) · 70.34 y Bhumibol of Thailand · 70.00 y Shapur II of Sasanian Persia
Headline
Sobhuza II 82.74 y > Louis XIV 72.30 y by 10.44 years (14 %), the largest single gap in the verified top-10

Next steps

  • Submit the top-7 data quality issues as Wikidata data-quality reports to fix the underlying records.
  • Repeat the query for related position classes (presidents, sultans, emperors as separate Q-codes) to extend the longest-tenure list across non-monarchic political offices.
  • Cross-validate Sobhuza II's reign length against the Eswatini royal archive for sub-day precision.
  • Investigate whether the 10.4-year gap between Sobhuza II and Louis XIV survives when you allow non-Q116 positions (e.g., paramount chiefs, traditional rulers) into the comparison.

Artifacts