Correcting the record.

How corrections are handled

We review reported errors against the source material. When a finding changes, the amendment identifies the edition, date, and wording affected. Readers receiving a corrected edition can distinguish it from the original.

Field Report 001, Ground Truth Is What Survives, was published for free public reading on October 21, 2026. Revision D, issued the same day, moves the report to its own web address with clearer prose and attribution, a linked contents list, and a redesigned reading layout. The evidence ledger, source links, and author credits are retained. Previously shared PDF addresses now lead to the web report; earlier files remain in the private editorial record.

Revision E, issued on October 21, 2026, shortens the research preview to protect nonpublic work. The investigation, evidence ledger, and source record are unchanged.

On October 21, 2026, reference links were updated to lead directly to the cited material. Reference 14 now gives the ICAO working paper’s cover date, September 10, 2025, in place of September 12. Reference 22 uses the publisher’s current article title. The report findings and evidence ledger are unchanged.

Revision F, issued on October 22, 2026, updates the research preview to identify hardware security bench work as the focus of follow-up reports. No bench findings are published in this revision. The investigation, evidence ledger, source links, and author credits are unchanged.

Interactive edition

Revision G, issued on October 23, 2026, adds source-linked confidence labels, key findings, an explanatory 3D scene, a coordinate map, evidence relationships, and a publication chronology. The scene is schematic. The map uses current reference geography with the investigator-provided coordinate. Source photographs, the evidence ledger, citations, author credits, and the location and munition findings remain unchanged.

OCTOBER 27, 2026 / FIELD REPORT 002 / REVISION B

Spatial reconstruction and investigator attribution

Revision B clarifies George Sudo (GeoSudoku)’s method in “Reconstructing the Road to Khaled bin Walid Camp.” He mentally reconstructed the road, terrain, buildings, parking area, and tower, then developed a sketch to test that layout against satellite imagery. The local place name led to a specific location, and the visual matches were found quickly. This case used spatial reconstruction, rather than triangulation or a brute-force candidate search.

GeoSudoku’s original investigation did not include AI-model geolocation tests, a 14-intersection / 12-road-cut / 3-approach elimination process, or GL4SS_HOUSE. The public web edition already omitted those claims; Revision B makes the method and authorship explicit. Any additional SSX360 technical analysis belongs in separately attributed sections with its own evidence.

The investigation and original geolocation are GeoSudoku’s. R. J. York prepared the web edition for SSX360. The eleven evidence figures, two source excerpts, reported coordinates, and limits concerning the filming date and control of the camp are retained. Triangulation examples from other investigations are not evidence for this case.

NOVEMBER 5, 2026 / FIELD REPORTS 001 AND 002

Article credits and method

R. J. York’s name is restored to the article credits: hardware and GNSS analysis and editorial work in Field Report 001, and SSX360 editorial preparation in Field Report 002. George Sudo (GeoSudoku) retains the investigation and geolocation credit in both reports.

Revision C of Field Report 002 refines the description of the method: “Mocha intersection” led to a specific location, and the reconstructed layout was checked against satellite imagery. It adds reading shortcuts to the working sketch and visual comparison. The eleven evidence figures, two source excerpts, reported coordinates, and findings are unchanged.

NOVEMBER 5, 2026 / FIELD REPORT 002 / REVISION D

Investigator account and source review

Revision D labels George Sudo (GeoSudoku)’s first-person account, adds his original Telegram investigation to the sources, and includes his Telegram and LinkedIn profiles. R. J. York’s SSX360 evidence review is separately credited. It discusses the publication record, the edit between video segments, and the distinction between observed structures and proposed outlines.

The superseded note about software testing is replaced by the attributed review. The eleven evidence figures, two source excerpts, reported coordinates, and location finding are retained. The filming date and control of the camp remain unresolved.

NOVEMBER 5, 2026 / FIELD REPORT 002 / REVISION E

Map comparison, terrain profiles, and reporting chronology

Revision E adds SSX360’s case-record checks, credited to R. J. York. Figure 12 compares GeoSudoku’s coordinates with an OpenStreetMap extract. Figure 13 plots SRTM and ASTER elevation profiles with separate sightlines under assumed camera and tower heights. Sources 6 to 12 document the map, elevation products, reporting chronology, and data-provenance limits.

Revision D presented the skyline object at about five seconds and the tower beside the camp gate at about twelve seconds as the same structure. SSX360’s terrain profiles place a ridge about 902 m from the camera, approximately 16 to 21 m above the modeled sightlines under the stated height assumptions. The same-tower identification is withdrawn; the fourth key finding now records it as not established. The location finding is unchanged and is also corroborated by the open map comparison. GeoSudoku reviewed SSX360’s profile, data, and conclusion and agreed to withdraw that identification. SSX360 subsequently corrected the curvature convention: the ASTER margin rounds to 16 m, rather than the earlier 15 m. The corrected calculation remains SSX360’s responsibility; no separate investigator review of that arithmetic is claimed. The original alignment discussion remains attributed to him and links to the later check. His first-person account, eleven source figures, two source excerpts, and reported coordinates are preserved.

The elevation figure uses each product’s own endpoint elevations and the effective-Earth-radius curvature convention. The received calculation is preserved beside the corrected output in the public data download. The reported margins are rounded from the saved model output. The figure does not claim a universal vertical-error bound, a measured tower height, or independent ground validation. Reporting dates are distinguished from filming dates; news claims of territorial control remain attributed to their sources.

The final provenance check establishes that the OSM camp boundary dates from April 13, 2017. Its 137 unique vertices are unchanged; the September 2026 edit changed only name tags. This replaces the earlier caveat about unestablished boundary history. Figure 13 and the direct JSON downloads use the corrected 0.13 m curvature term and 20.94 m / 15.65 m margins. These presentation and provenance updates change no finding.

NOVEMBER 6, 2026 / FIELD REPORT 002 / REVISION F

Height sensitivity and final assessment

Revision F adds SSX360’s height-sensitivity calculation from the same 100-point terrain profiles. With the camera fixed at 2 m, the modeled tower thresholds are 86.5 m for SRTM and 68.4 m for ASTER. With the tower fixed at 15 m, the camera thresholds are 31.6 m and 24.1 m. These total heights and the new JSON record are calculated results, not field measurements. The saved 21 m / 16 m obstruction margins and Figures 12 and 13 are unchanged.

The final assessment records the same-tower identification as not supported by the saved model. It retains the documented September 11 release as the latest possible filming date, while distinguishing that bound from an unverified post-2023 lower bound and a narration-dependent September 9 to 11 window. Contemporary reports of the camp’s capture remain attributed; no uninterrupted-control or universally undisputed-capture claim is made.

GeoSudoku’s location finding, first-person account, source figures, videos, and coordinates are unchanged. The new calculation and final assessment are SSX360’s work under R. J. York’s byline. GeoSudoku’s prior review remains attached to the earlier terrain finding and is not represented as approval of the new analysis.

Earlier correction history

AUGUST 11, 2026 / EARLIER REPORT SERIES

August Report 001 republished

Report 001 was republished with a new subject, finding, source record, and named author. It replaced the August 9 edition on framework counts and effective dates. Both editions remain in the private editorial record.

For a correction or citation inquiry, identify the edition in your message. Do not include confidential source material in an initial email.