Corrections

As published. On the September 2026 France report the row (Signal | Read | So what for your supply chain) read 'Transportation weakest at 40.0 | principal drag | the main brake on the composite — the only sub-index below 50', while nine of its ten sub-indexes were below 50. In September the 'only' wording was wrong on 33 of the 42 country reports, and 'the main brake on the composite' on 18: on 6 the weakest sub-index was a zero-weight diagnostic (Logistics Cost or Inventory), on 8 another weighted sub-index weighed more on the composite, and on 4 no weighted sub-index was below 50. Across the January–September editions the figures are 290 and 148 of 378 country reports.

Corrected. The France row now reads 'Transportation weakest at 40.0 | principal drag | the main brake on the composite — 9 of 10 sub-indexes are below 50'. Every country report now states how many sub-indexes are below 50, or that none is; calls the weakest sub-index the main brake only when it has the largest negative weight × distance below 50 (computed from the published figures, so an exact tie is a joint brake), and otherwise names the sub-index that has; says a zero-weight diagnostic does not move the composite, naming the weighted brake beside it where there is one; and describes a weakest sub-index shown at 50.0 as at the no-change line.

Basis. WRI editorial defect in the report template's Signal watch, present since the site launched on 24 June 2026 and shown on country reports of every edition: the 'only' wording was not conditioned on the count, and the brake was taken from the lowest level rather than from weight × distance below 50. No index value, weight or published reading changes. Found by WRI's own review on 9 October 2026.

As published. On the September 2026 World report the line read 'Above the 50 no-change line; an even split of accelerating and slowing — 8 of 10 in expansion', directly above tiles showing Accelerating 5, Slowing 3 and Deteriorating 2. The same template produced every Logistics Index and Logistics Growth Index (LGI) report from the February 2026 edition, and its counts did not follow the published one-decimal figures: the September America report gave '2 rising, 8 easing this month' although Transportation was unchanged (0.0); the June 2026 Europe report gave '5 rising, 5 easing this month' and 'an even split of accelerating and slowing', counting Goods for Sale, unchanged at 0.0, as rising; and the May 2026 Europe report gave '6 of 10 sub-indexes in expansion, 3 at the no-change line', counting two sub-indexes shown at 50.0 both ways. On country reports the Signal watch called any reading within 0.8 of 50 'at the 50 line' and an unchanged one 'tilting down' (United Kingdom, February 2026: 'Employment at the 50 line (49.5, 0.0) | tilting down | sinking further below 50').

Corrected. The World report's line now reads 'Above the 50 no-change line; as many sub-indexes rising as easing (5 each) — 8 of 10 in expansion; 6 of the 8 weighted.' Every report now counts each sub-index as published, to one decimal: a move shown as 0.0 is unchanged, neither rising nor easing (America, September: 2 rising, 7 easing, 1 unchanged; Europe, June 2026: 4 rising, 5 easing, 1 unchanged), and a reading shown as 50.0 is at the no-change line, not in expansion, which changes one published count (Europe, May 2026: 4 of 10 sub-indexes in expansion, not 6). Sub-indexes shown at 50.0, which the tiles left out, and sub-indexes unchanged on the month, which were filed under a rising or falling group, are now named under the tiles. On country reports the Signal watch now says 'near the 50 line' and calls an unchanged sub-index flat (United Kingdom, February 2026: 'Employment near the 50 line (49.5, 0.0) | flat on the month | unchanged on the month — the next print sets the direction'). The zero-weight diagnostics (Logistics Cost and Inventory) are labelled wherever the report template names them; the written commentary is unchanged. Each report's breadth line and breadth tile add the count among the 8 weighted sub-indexes, and from the October 2026 edition breadth counts only those 8.

Basis. WRI editorial defect at publication in the report template, which has produced every report since the February 2026 edition: the sentence compared rising with falling sub-indexes but used the names of the above-50 groups, and the breadth counts and the Signal watch did not treat sub-indexes at the published one-decimal precision. No index value, weight or published reading changes. Found by WRI's own review on 9 October 2026.

As published. Each World release page took the name of its previous month from the latest month in the series instead of from its own month, so the name moved on with every new edition. After the August edition's release on 10 September, for example, the July page read 'The World reading registered 51.9, +0.3 on 51.6 in July 2026', where 51.6 is the June reading. The readings and changes were correct; only the month name was wrong. Country release pages were not affected.

Corrected. Each World release page now names the month before its own: the July page reads 'The World reading registered 51.9, +0.3 on 51.6 in June 2026'.

Basis. WRI template defect, present since the release pages launched on 19 August 2026: the World page took its previous month from the second-to-last month of the whole series rather than from the month before the page's own. Found by WRI's review of its public pages on 6 October 2026 and corrected the same day.

Before. The August 2026 edition and its release pages stated that the September 2026 edition publishes on 10 October 2026, 09:00 UTC — accurate under the schedule then in force (editions released on the 10th of the month after the reading month).

Updated. Update, not correction: the publication schedule changed by decision of 23 September 2026 — from the September 2026 edition the Index publishes on the 7th of each month at 09:00 UTC, regardless of weekday; the September edition follows on 7 October 2026. Historical editions keep their true release dates (the 10th), including in datelines and structured data. No reading, methodology or computation changes.

Basis. Schedule change by WRI decision (announced and dated; release-date computation is edition-aware from the September 2026 edition). The Gateway Access Monitor's separate schedule — the 1st of each month from its October 2026 edition — is unaffected.

As published. The August edition release page listed 'United Arab Emirates — largest rise, +1.5 on the month' while the edition's own CSV shows South Korea at +1.8 (South Korea appeared under its 'highest of 42' story instead), and both Russia (−2.6) and South Africa (−2.3) were labelled 'sharpest fall'.

Corrected. Story labels now reflect true rank positions regardless of which story slot a country occupies: the United Arab Emirates reads 'second-largest rise, +1.5' and South Africa 'second-sharpest fall, −2.3'; South Korea's +1.8 remains the month's largest rise, told through its 'highest of 42' story. All readings, changes and ranks are unchanged.

Basis. WRI labelling defect at publication: the one-story-per-country rule let a runner-up inherit a superlative label. Caught by the 10 September outreach wave review (verification against the edition CSV) and corrected the same day; labels are now assigned by true rank position.

As published. The Brazil and Norway releases described August's readings as the lowest in the published series, where at published one-decimal precision each ties an earlier published reading (Brazil 50.5 ties January 2026; Norway ties its earlier low). The Israel release said 'lowest reading in 3 months' where the reading two months prior ties at one decimal, and the Italy release said 'highest reading in 2 months' on a rise invisible at one decimal (a 0.0 displayed change).

Corrected. Superlative claims are now computed at the published one-decimal precision every reader can check against the open CSVs, with ties stated as such: Brazil and Norway read 'joint-lowest reading in the published series', Israel 'lowest reading in 2 months', and Italy carries no superlative. The same recompute upgrades Indonesia's true '5-month high' to the stronger 'joint-highest reading in the published series' (August ties March 2026). All index readings, changes and ranks are unchanged.

Basis. WRI computation defect at publication: record and window comparisons ran on stored full-precision values rather than the published one-decimal figures. Caught by the 10 September outreach wave review (the Brazil draft's 'series low' failed verification against the public CSV) and corrected the same day; the recompute is now the guard for every release surface.

As published. The July edition displayed Europe at 49.8, Switzerland at 52.7 and the United Kingdom at 49.2. Each figure had been rounded twice: the full-precision index value was stored at two decimals in the source workbook and then rounded again to one decimal for display, which pushed these three half-boundary cases up by 0.1.

Corrected. Displayed readings are now rounded directly from the full-precision index values: Europe 49.7 (from 49.7477), Switzerland 52.6 (from 52.647), United Kingdom 49.1 (from 49.145). A systematic sweep across every published reading found no other affected display. Underlying values, month-on-month changes at full precision, ranks and trend classifications are unchanged.

Basis. WRI presentation artefact at publication, removed by the 9 September precision upgrade (index values now published from full working precision). Surfaced by an external review on 10 September and disclosed the same day.

As published. The July World report stated that Logistics Cost "led the move (+1.2)" and "added the most to the move". Logistics Cost carries a 0.00 composite weight — its weighted contribution to the headline is zero — so the sentences conflated a diagnostic's standalone movement with an index contribution.

Corrected. Driver sentences now rank weighted sub-indexes by weight-times-change: for July, Warehousing led the move (+0.7), with Transport Capacity the main drag (-0.8), and the Logistics Cost diagnostic is reported separately as the largest single move (+1.2) at zero composite weight. The correction applies to every report generated by the same logic.

Basis. WRI editorial defect at publication: the report generator selected "movers" by raw month-on-month change across all sub-indexes, weighted and diagnostic alike. All index values, weights and the published headline are unchanged; only the attribution wording was wrong. Flagged by an external review on 8 September; corrected the same day.

As published. Every May 2026 report said '32 of 42 economies are in expansion this month', counting Austria (50.02) and the United Kingdom (50.01), both published as 50.0. The 'Months above 50' box counted months published at 50.0 that were at or fractionally above 50: Europe's September 2026 report gave '3 of 9' and Africa's '6 of 9', counting Europe's January reading of 50.0 and Africa's July reading of 50.03; China's reports from February 2026 counted its January 50.0, and Austria's and the United Kingdom's from May 2026 counted their May readings. On country reports the Decision read called a headline published at 50.0 'expanding' or 'contracting' (Austria, May 2026: 'Net — Austria is expanding and slowing'; China, August 2026, 49.96: 'Net — China is contracting and deteriorating'), and China's forecast record counted its August outcome of 50.0 as an expansion hit ('Direction hit 4 / 4' in September).

Corrected. May 2026 reports now say 30 of 42 economies are in expansion; Austria and the United Kingdom, shown at 50.0, are not counted. The 'Months above 50' box counts only months published above 50.0 (Europe, September 2026: 2 of 9; Africa, September 2026: 5 of 9). The Decision read describes a headline published at 50.0 as at the no-change line (Austria, May 2026: 'Net — Austria is at the no-change line and edging down'), the header says 'fell', not 'eased', for such a headline, and China's forecast record reads 'Direction hit 3 / 4' in September (2 / 3 in August). 73 reports changed in all.

Basis. WRI editorial defect at publication in the report template: these counts and statements compared unrounded readings with 50, not the published one-decimal figures — the rule the breadth counts adopted earlier the same day. No index value, weight or published reading changes. Found by WRI's own review on 9 October 2026.

Before. At release the Panama outlook listed the ACP's September operations summary as a dated input still to land (~10 October, carrying September Canal Waters Time and the culvert outage's completion status). Until this update the watch line on the Miraflores west-lane culvert outage read “no advisory confirms, postpones or completes it”, and the waiting-time figure gave August Canal Waters Time (34.90 h) as the latest monthly authority measure. The page's Panama traffic figures are the IMF PortWatch ship-tracking (AIS) series — 28.1 transits a day in September, 10% below its 2025 average — which the page names as a different population from the ACP's own count.

Updated. Update, not correction — a document issued after publication. A dated note now closes the Panama outlook: “Update — 10 October. The ACP's September operations summary (A-38-2026, 9 October) puts the canal's own count at 32.6 transits a day, about level with its 2025 average of 33.1. The ship-tracking (AIS) series this page uses at every chokepoint reads 28.1 a day, 10% below its own 2025 average; since August it has run at a smaller share of the canal's count (87% in August and 86% in September, against 90–96% in every month from March 2024 to July 2026), for reasons not yet known. The summary also puts average Canal Waters Time at 37.97 hours, up from 34.90 in August (monthly high 78.15), and lists the Miraflores west-lane culvert outage of 22 September–3 October as completed; this edition's waiting-time direction (deteriorating, on August's measure) stands, and the September figures are read in the November edition. No figure or direction changes.”

Basis. ACP Advisory to Shipping A-38-2026, Monthly Canal Operations Summary – September 2026 (9 October 2026): 979 oceangoing transits, a daily average of 32.63 (high 37, low 28) — 1.5% below 33.12 a day, WRI's mean of the ACP's twelve 2025 monthly averages, inside the 5% band the page applies to monthly transit means, hence “about level”; average Canal Waters Time 37.97 h (high 78.15, low 17.60); its Panamax maintenance schedule lists the Miraflores west culvert outage, 22 September–3 October 2026, as “Completed”. IMF PortWatch's full-September vintage (6 October) gives Panama 28.13 transit calls a day; as a share of the ACP's count that series ran at 0.900–0.963 in every month from March 2024 to July 2026 (0.864 in January 2024; the ACP published no February 2024 summary), 0.873 in August 2026 and 0.862 in September 2026 (WRI's comparison of the two published series, 9–10 October 2026).

As published. The Suez outlook read (as refreshed on 8 October) “On the daily all-vessel series this edition reads, September (40.4/day, full month) sits 1.5% below August …”, “The container return is visible in carrier and analyst counts rather than in the all-vessel mean …” and “(press; a different population from the all-vessel series)”.

Corrected. It now reads “On the daily PortWatch series (five ship types) this edition reads, September (40.4/day, full month) sits 1.5% below August …”, “… rather than in the PortWatch mean (five ship types) …” and “(press; a different population from the PortWatch series)”.

Basis. WRI editorial defect at publication: IMF PortWatch's daily transit calls are not an all-vessel count — its description of the daily chokepoint dataset says “all the indicators are available by 5 main ship categories: container, dry bulk, general cargo, ro-ro and tanker”, and its daily total is the sum of those five. Found in WRI's check of the published editions on 9 October 2026.

Before. At release the September figures were provisional, from IMF PortWatch daily AIS calls to 27 September (27 of 30 days): Panama 28.2 transits a day, Suez 40.7, Bab el-Mandeb 26.2, Strait of Hormuz 3.7, Cape of Good Hope 86.6, Gibraltar 133.3 and Malacca 219.4, with the Strait of Hormuz's 7-day rolling mean below 80% of its August mean on 15 of 27 days.

Updated. Scheduled data finalisation under methodology v1.1 §9, not a correction. On the full-September vintage the September means are Panama 28.1 (3.1% below August, inside the 5% band), Suez 40.4 (1.5% below), Bab el-Mandeb 26.6 (2.9% below), Strait of Hormuz 3.6 (23% below; its 7-day rolling mean under 80% of the August mean on 18 of 30 days), Cape of Good Hope 86.1, Gibraltar 133.1 and Malacca 219.3. Against their references: Panama −10% (was −9%), Bab el-Mandeb −65% (−66%), Gibraltar −11% (−10%) and Cape of Good Hope +67% (+68%); Suez, the Strait of Hormuz and Malacca are unchanged at −46%, −96% and 0%. Every direction call stands, and the provisional labels are removed. The Strait of Hormuz “In brief” line, which read “15 of 27 September days below the August level”, now names the threshold it counts: “18 of 30 September days below 80% of the August level”. The edition's evidence date stays 1 October; the page now adds the date the ship-traffic figures were finalised.

Basis. Methodology v1.1 §9: an edition's prior-month readings publish provisional and are finalised by a dated data refresh when the full-month PortWatch vintage lands. Annex B: of the 406 observation-days carried in both vintages (seven gateways, 1 August–27 September), none was revised; 28–30 September were added. How the final vintage could move a direction call was set down in writing on 6 October, before the data arrived; no call crossed the 5% band.

Before. The Panama recovery-flexibility figure read “the late-cancellation surcharge postponed by A-32 (2 September) has no new date”; the recovery-flexibility direction row read “A-32's postponed late-cancellation surcharge has no new date”; the announced-measures figure read “The live advisory index listed nothing after A-36 at 06:15 UTC on 1 October”, and the sources note “none after A-36” — accurate when the edition was released on 1 October.

Updated. Update, not correction — a document issued after publication: ACP Advisory A-37-2026 (5 October) applies the Booking Late Cancellation Surcharge (tariff 1050.IBC9, announced in A-28-2026) from 15 October until further notice, at 250% of the reservation fee in addition to the cancellation charges, to cancellations received four days or less before the required arrival time and to vessels that fail to arrive within seven days after that time. A-37 describes the change as a reduction from the seven days announced in A-28; its own table lists the current notice period as two days and the new one as four. A-37 does not refer to A-32 (2 September), which postponed the A-28 measure until further notice. A-37 also offers long-term slot packages for transit dates 3 January–3 April 2027: a Neopanamax competition on 17 November and, for the first time, advance Panamax slots by sealed-bid auction on 24 November, each averaging three slots a day. The two figures, the direction row and the sources note now say so. A-37 changes none of the edition's readings, directions or access states; the November edition reads the new terms.

Basis. Evidence after publication (methodology v1.1 §9: a change caused by a document issued after publication is recorded as an update). Sources: ACP Advisory to Shipping A-37-2026 (5 October 2026; primary); ACP advisory index, re-listed 00:00 UTC on 8 October.

As published. The waiting-time figure read “The latest authority measure, August Canal Waters Time (A-34), reads 34.90 h …” and “The ACP publishes no queue length between monthly summaries; the September summary (~10 October) is the next measure.”; the waiting-time direction row read “The authority's latest published measure, August Canal Waters Time 34.90 h (A-34) …”; and the transparency scorecard gave the ACP's publications as “Monthly PDF” (transits) and “CWT monthly” (waiting).

Corrected. The figure now reads “The latest monthly authority measure, August Canal Waters Time (A-34), reads 34.90 h …” and “Between its monthly summaries the ACP publishes a daily count of the vessels in queue at 00:01, by lock class, in its daily customer report; this edition did not use it. The September summary (~10 October) is the next monthly Canal Waters Time measure.”; the direction row reads “The authority's latest monthly measure, …”; and the scorecard cells read “Monthly PDF; daily report” and “CWT monthly; queue and CWT daily”.

Basis. WRI editorial defect at publication: the ACP's Enhanced Vessel Traffic Management System publishes an Operations Department Daily Customer Report whose section “D. YESTERDAY” gives actual arrivals, actual transits and an average Canal Waters Time by lock class, and whose section “E. QUEUE” gives the vessels in queue “AT 0001”; these reports were public before this edition was released. Found in WRI's review of the Panama sources on 7 October 2026.

As published. Access runs through Iran's Persian Gulf Strait Authority (PGSA) case-by-case permits (no reservation product) and parallel US-facilitated passage on the southern, Omani-side route coordinated by US Naval Forces Central Command (NCAGS).

Corrected. Access runs through Iran's Persian Gulf Strait Authority (PGSA) case-by-case permits (no reservation product) and parallel US-facilitated passage on the southern, Omani-side route coordinated through US Naval Forces Central Command's Naval Cooperation and Guidance for Shipping (NCAGS). The rest of the figure, the allocation-rules reading and the Hormuz direction count are unchanged.

Basis. WRI editorial defect at publication: the abbreviation was expanded as the name of the command. NCAGS is the Naval Cooperation and Guidance for Shipping function of US Naval Forces Central Command (NAVCENT), as the Joint Maritime Information Center's advisory notes relayed by UKMTO style it ('NAVCENT NCAGS'). Found in WRI's readability review of the page on 6 October 2026.

As published. The Suez outlook read “… — but on the daily all-vessel series this edition reads, the 7-day rolling mean held at or above the July mean for at most seven consecutive days (21–27 August) …”.

Corrected. It now reads “… — but on the daily PortWatch series (five ship types) this edition reads, the 7-day rolling mean held at or above the July mean for at most seven consecutive days (21–27 August) …”.

Basis. WRI editorial defect at publication: IMF PortWatch's daily transit calls are not an all-vessel count — its description of the daily chokepoint dataset says “all the indicators are available by 5 main ship categories: container, dry bulk, general cargo, ro-ro and tanker”, and its daily total is the sum of those five. Found in WRI's check of the published editions on 9 October 2026; the same wording is corrected in the October 2026 edition.

As published. The waiting-time figure read “The ACP publishes no queue length between monthly summaries.”; the waiting-time direction row read “No new authority figure at the data cut — the ACP's August operations summary follows in the October edition; July CWT 25.4h stands as the latest published measure.”; and the transparency scorecard gave the ACP's publications as “Monthly PDF” (transits) and “CWT monthly” (waiting).

Corrected. The figure now reads “Between its monthly summaries the ACP publishes a daily count of the vessels in queue at 00:01, by lock class, in its daily customer report; this edition did not use it.”; the direction row reads “No new monthly authority figure at the data cut — … July CWT 25.4h stands as the latest monthly measure.”; and the scorecard cells read “Monthly PDF; daily report” and “CWT monthly; queue and CWT daily”.

Basis. WRI editorial defect at publication: the ACP's Enhanced Vessel Traffic Management System publishes an Operations Department Daily Customer Report whose section “D. YESTERDAY” gives actual arrivals, actual transits and an average Canal Waters Time by lock class, and whose section “E. QUEUE” gives the vessels in queue “AT 0001”; these reports were public before this edition was released. Found in WRI's review of the Panama sources on 7 October 2026; the same statements are corrected in the October 2026 edition.

Before. The 23 September update carried the strike sequence through Warnings 140-26 and 141-26 with crews safe throughout the tracked window, the IMO count at 80 attacks on the 15 September list, the diplomacy position as an offer publicly rebuffed the same day, and MARAD 2026-006 past its 22 September expiry with no successor posted.

Updated. Update, not correction — documents and evidence issued after publication: UKMTO Warning 142-26 (report time 06:30 UTC, 23 September; UKMTO's wording is “2 casualties reported” and its warning does not name the vessel); Oman Maritime Security Centre's statement (Cape Dao named, 27 crew evacuated by the Royal Navy of Oman, engine-room fire, about 2.5 nautical miles off Musandam); and the Embassy of India in Muscat's confirmation of one Indian national killed among 19 aboard — the first fatality among attack-classified strait incidents in the tracked sequence, with crew figures recorded per holder, no claim of responsibility, and the vessel's status held open (last authoritative state: on fire and adrift). The IMO's own page states 84 confirmed incidents and 23 confirmed seafarer fatalities as at 23 September, with listed entries through 21 September and the Cape Dao not yet included — the records are not summed. A completed round of indirect, mediated US–Iran talks in New York on 22 September, both principals on the record and the strait's reopening on the agenda, is recorded on the outlook strand only as a stage change from the offer-and-rebuff position. MARAD Advisory 2026-013 (effective 23 September, expiring 22 March 2027) supersedes and cancels 2026-006, carrying the first US-authority placement of the reported 24 August strikes on the tankers Amzan and Daisy off Yanbu. Visible Yanbu loadings remained at zero at the 24 September re-check. The edition's readings, directions and access states are unchanged.

Basis. Evidence after publication (Methodology v1.0: later documents and data generate a dated update, never a silent edit). Sources: UKMTO live incident feed (primary; the warning PDF unposted at retrieval), Oman Maritime Security Centre statement (authority), Embassy of India in Muscat statement via press carries (authority-class, quoted), IMO highlighted-incidents page (authority, 23 September vintage), MARAD advisory pages (authority), and press accounts of the New York talks carrying both principals' on-record statements (press; the direct-versus-indirect characterisation is contested and flagged). Union statements are recorded as union-class claims and not carried as fact.

As published. The 23 September update read: “restoration to the 4 million barrel-per-day reroute level estimated at six to eight weeks”. The Reuters wire of 22 September states the 4 million bpd target (“Aramco was seeking to get the pumping rate back to 4 million bpd, one of the sources said”) and, separately, “Reaching a rate of 40% of capacity will take a couple of days and a full restart will take 6 to 8 weeks, a security source said”, with a third source adding “a return to full pumping rates would take up to six weeks”.

Corrected. The update now reads: “Aramco seeking to return the rate to the 4 million barrel-per-day reroute level (nameplate capacity 7 million). Full-restoration timing is uncertain and level-unspecified in the sourcing: a security source estimated roughly 40% of capacity within days and a full restart in six to eight weeks, while a separate industry source put a return to full pumping rates at up to six weeks.” Bloomberg's independent reporting attaches its roughly-six-week horizon to full capability, consistent with the decoupling. The restart itself remains recorded as a claim, with visible Yanbu loadings at zero since 16 September and the suspension standing.

Basis. WRI wording defect in the dated update: two separately attributed statements in the source wire were merged, and a competing estimate omitted. Flagged by an external review on 24 September, verified against the wire verbatim at two independent carries, and corrected the same day (Research Standards: corrections are errors at publication, distinct from updates). The direction of the error was pessimistic — binding the longer estimate to the lower level.

Before. The 21 September update carried the strike sequence through the two unnumbered 18 September entries, the pipeline at three damaged stations with a five-to-six-week repair estimate and the Yanbu suspension in force, MARAD 2026-006 as the standing Red Sea advisory, the Gemini return as two Suez-bound broadcasts, and the comparable table on the PortWatch vintage running to 13 September.

Updated. Update, not correction — evidence and documents issued after publication: UKMTO Warnings 140-26 and 141-26 (21 September; vessels identified by press as LR Stephanie and Al Maryah — names not carried by UKMTO's products, incident and report dates recorded separately; the coalition's JMIC has meanwhile corroborated the Trend's identity for 137-26 and named STI Steadfast for 138-26); Reuters' reported pipeline restart at a low pumping rate (three of eleven pumping stations damaged, six-to-eight-week full restoration estimate, one China-bound cargo scheduled at Yanbu) beside Bloomberg's early-stage start-up characterisation targeting meaningful flows by 26 September — a claim, with visible Yanbu loadings at zero since 16 September and the suspension standing; MARAD 2026-006 past expiry unreplaced; the Suez Canal Authority's confirmation of OOCL Portugal's 16 September southbound transit (Bab el-Mandeb passage not yet observed and held open); Cornelia Maersk's departure from Colombo two days ahead of schedule and Maastricht Maersk's departure via Tanger Med, completing four Suez-declared first-wave vessels; Monday's visible transit print of two (a floor, not an exhaustive record) and Sunday's three-vintage revision 6 → 7 → 10; Gulf-terminal loadings of 12–14 million barrels on Sunday recorded as a range across contested tracker counts; and Kpler's record 7.2 million barrel-per-day September Gulf-of-Oman ship-to-ship estimate with its own double-counting caveat. The edition's readings, directions and access states are unchanged.

Basis. Evidence after publication (Methodology v1.0: later documents and data generate a dated update, never a silent edit). Sources: UKMTO warning PDFs (primary), Marisks via Reuters (press vessel identifications, confidence-labelled), Reuters and Bloomberg wires (pipeline, independent sourcing, characterisations carried side by side), MARAD portal (authority), Suez Canal Authority release (authority) with tracker corroboration, AIS aggregators (declarations, single signal class), TankerTrackers.com via Reuters and MarineTraffic via press (contested range), Kpler analysis (measured volumes; capacity lines Kpler-attributed), IMF PortWatch (comparable series, vintage to 20 September).

Before. The 16 September update read: Reuters reports Gulf producers sustaining 7–9 million barrels per day — 30–40% of pre-war volumes — via so-called dark shipments. An internal review (Insights desk, 20 September) flagged that this could be read as a total-flows claim in tension with Reuters' 9 September data piece.

Updated. Update, not correction — the sentence is a near-verbatim, correctly attributed relay of the Reuters wire's own wording (verified against the wire text by two independent checks: the 30–40% clause is the agency's, attached to the dark-shipment figure; the adjacent trading-sources tag governs only the preceding sentence). The tension between the wire and Reuters' 9 September analysis is between two Reuters publications, measuring different populations — dark-enabled export volumes versus total Gulf flows including dark crossings. A dated scope note now records the distinction on the page; no reading, direction or access state changes.

Basis. Internal review flag (Insights desk disposition, 20 September) resolved under the proven-mechanism rule: no error mechanism on the Monitor's side exists, so no correction; the ambiguity is source-side and is now documented rather than silently inherited. Sources: Reuters wire, 15–16 September (verified via syndicated carry); Reuters data analysis, 9 September (Goldman Sachs, Vortexa; verified via syndicated carry).

Before. The 16 September update carried Warning 135-26 as the latest strike report, Monday's series-low first print of four, the three-to-five-week pipeline repair estimate (15 September, AP-attributed), the Yanbu loading suspension, and the edition's A-31 booking rules with no A-35 reference; the Gemini return to Suez stood as announced schedules only.

Updated. Update, not correction — evidence and documents issued after publication: UKMTO Warning 137-26 (17 September, Khasab) plus two 18 September Attack entries without public designation numbers (one onboard fire extinguished; one time-late 16 September outbound strike; crews safe); the IRGC Navy's claim of striking the tanker Trend for transit outside Iran's approved-passage system (Iranian state media 17–18 September; AP could not confirm; no UKMTO product names the vessel); Kpler daily-count revisions 4→7, 4→12 and 3→6 (fifteen of sixteen observation-days upward); Maersk operational update 48 (17 September: USD 1,800/3,000/3,800 emergency freight plus USD 1,000 per container for Hormuz transits, replacing landbridge charges); Reuters trade-source reporting of about 60 million barrels sold for September–October loading via ship-to-ship transfer at Sohar, with the Petroleum Association of Japan on the record on risk allocation; three damaged pumping stations and a five-to-six-week full-repair estimate with a press-reported half-capacity bypass target; the Saudi-led coalition's statement that attempted attacks including on Yanbu were thwarted (19 September; Houthi claims of strikes carried no evidence and no damage was independently established); Italy's national-command escort activation for Bab el-Mandeb (17 September); JMIC's 13 September advisory formally listing Mokha, Dhubab, Perim and the Hanish Islands as Houthi-held and recording about 70 US-facilitated transits over 10–12 September; AIS-observed departure of Antonia Maersk from Tanjung Pelepas (19 September 08:40 UTC, westbound leg not yet observed) with the carriers' notices differing on one sailing date (vessel movements support 24 September for the Colombo departure); Drewry's WCI at USD 4,500 with Asia–Europe rates falling on the Suez return; and ACP Advisory A-35-2026 (16 September), a loosening clarification of Panamax consecutive-date booking within the unchanged 32-slot regime. The edition's readings, directions and access states are unchanged.

Basis. Evidence after publication (Methodology v1.0: later documents and data generate a dated update, never a silent edit). Sources: UKMTO incident listing (primary), IRGC statement via Iranian state media with AP relay (attributed belligerent claim), Kpler counts via Reuters (preliminary, revisable), Maersk operational update 48 (carrier primary), Reuters trade sources with a Petroleum Association of Japan on-record statement, satellite-analysis and industry-source reporting via Reuters and Kpler, Saudi-led coalition and Houthi statements via AP (attributed claims, both sides), ANSA/Reuters/LaPresse for the Italian activation, JMIC Update 096 (primary), AIS aggregators (single signal class), Drewry WCI (primary), ACP Advisory A-35-2026 (primary).

Before. The edition stated that the October 2026 edition publishes on 10 October 2026 — accurate when published under the schedule then in force — and its 15 September update carried the Yanbu stock-cover estimate, Warning 134-26 as the latest strike report, and Sunday's Bab el-Mandeb first print of 27.

Updated. Update, not correction: the publication schedule changed by decision of 16 September — from October the Monitor publishes on the 1st of each month at 09:00 UTC; the October edition follows on 1 October 2026, reporting September readings provisional at release with a dated final data refresh when the full-month PortWatch vintage lands. Methodology v1.1 formalises the cadence and the two-vintage model. New evidence recorded the same day: oil loadings at Yanbu are suspended and some September-loading cargoes cancelled to European customers (Reuters, trade sources — on roughly day five of the shutdown, consistent with the attributed five-to-seven-day stock cover); UKMTO Warning 135-26 (report dated 14 September, issued 15 September) records another vessel struck, with no damage reported; visible Hormuz transits printed four on Monday 14 September, the series low; Sunday's Bab el-Mandeb count was revised 27 → 28, as the page's first-print caveat anticipates; and Reuters reports Gulf producers sustaining 7–9 million barrels per day (30–40% of pre-war volumes) via dark shipments, with Vortexa recording a sharp Saudi Gulf-side loading surge. The edition's readings, directions and access states are unchanged.

Basis. Schedule change by WRI decision (announced, dated, and formalised at methodology v1.1); evidence after publication for the remainder (Methodology v1.0: later documents and data generate a dated update, never a silent edit). Sources: Reuters (trade sources; LSEG pricing), Vortexa via Reuters, UKMTO Warning 135-26 (primary), Kpler counts via press (labelled preliminary and revisable).

As published. The 15 September update's Panama note (and the matching register entry) read “a 37.5% increase consistent with this edition's deteriorating waiting-time call”. The September edition's direction table classifies Panama Waiting time as stable — “No new authority figure at the data cut — the ACP's August operations summary follows in the October edition” — while Cost of secured passage is the dimension classified deteriorating.

Corrected. The update now reads: “a 37.5% increase; the edition's waiting-time direction, stable on the information available at its cut, is unchanged, and the August figures inform the October reassessment.” The A-34 figures themselves are unchanged and still feed the October reassessment; the September classification stays frozen at its published value, as the methodology requires.

Basis. WRI wording defect in the dated update: the trend label of the adjacent dimension row was misattributed to Waiting time when the update was drafted. Flagged by an external review on 16 September and corrected the same day; direction calls are read from the published direction table, which was and remains correct.

Before. The 11–12 September updates carried the islands' status as unresolved on single-side sourcing, the 10 September full-day traffic recovery, and the pipeline shutdown without a repair horizon; the edition carried July as the latest ACP waiting-time measure, with the August summary expected at the October edition, and MSC's surcharges as announced but not yet in force.

Updated. Update, not correction: AP reporting sourced to both government and Houthi officials records Perim (Mayun) held and Greater and Lesser Hanish — about 160 km north of the strait — captured. Bab el-Mandeb transits ran 15 on 11 September (the day Perim was reached), then 24 and 27 over the weekend against a 10-day average near 27 (Kpler via Reuters, preliminary; first prints on this series have so far revised upward) — the level recovered while the security regime changed. In Hormuz, UKMTO Warning 134-26 records a vessel struck at 23:00 UTC on 12 September; Iranian state media report one crew death and four injuries aboard an Iranian container vessel near Hengam Island the same night; the Financial Times reports US air-defence protection narrowed to two daily time slots after night attacks increased; weekend traffic held in single digits against a 10-day average of 14; and the Salalah meeting of 14 September was postponed. The East-West pipeline repair is estimated at three to five weeks (two regional officials briefed, per AP; partial operation possible) against attributed five-to-seven-day Yanbu stock cover (Reuters). MSC's Piracy Risk (US$55/TEU) and Suez Canal (US$36/TEU) surcharges entered application on 15 September by proforma date. The ACP's August operations summary (A-34, posted 10 September 15:46 UTC, after the edition's release) shows average Canal Waters Time at 34.90 hours against July's 25.38 — consistent with the edition's deteriorating waiting-time direction — and is incorporated fully at the October edition. The edition's readings, directions and access states are unchanged.

Basis. Evidence after publication (Methodology v1.0: later documents and data generate a dated update, never a silent edit). Sources: AP (via wire syndications), Reuters/Kpler daily counts (labelled preliminary and revisable), UKMTO Warning 134-26, Iranian state media (attributed), Financial Times via Reuters, Omani foreign ministry statement, MSC customer advisory, ACP Advisory A-34-2026 (server-dated 10 September 15:46 UTC).

Update, 16 September 2026. This entry's Panama sentence (“consistent with this edition's deteriorating waiting-time call”) is corrected by the 16 September log entry: the edition's waiting-time direction is stable; “deteriorating” is the Cost of secured passage row. The A-34 figures and everything else in this entry are unchanged.

Before. The 11 September update recorded the capture of Mocha, preliminary intraday tracking showing Bab el-Mandeb crossings falling sharply on 10 September, and no post-capture UKMTO or JMIC product.

Updated. Update, not correction: Reuters reported on 11 September, citing four Yemeni government sources, that Houthi forces had reached Perim (Mayun) island inside the Bab el-Mandeb strait — two of the sources said government forces had withdrawn from the island — and had taken the mainland coastal town of Dhubab facing it (single-side sourcing on the island's status; no closure of the strait is reported). Saudi Arabia separately shut the East-West pipeline, the overland Hormuz bypass feeding Red Sea loadings at Yanbu, as a stated precautionary measure after drone strikes on pumping stations in the Riyadh and Medina regions on 10 September (Saudi energy ministry via wires; drones attributed by both governments to launches from Iraq; no restart date or export-loss figure published). Full-day ship-tracking for 10 September — the capture day — shows 26 Bab el-Mandeb transits against a 10-day average near 27 (Kpler via Reuters, preliminary): the sharp intraday fall recorded in the 11 September update did not persist across the full day. The same vintage revised Hormuz transits for 9 September upward from seven to eleven, and the QatarEnergy carrier Al Marrouna delivered a Ras Laffan cargo to Pakistan on 10 September — the first known Qatar-linked LNG shipment since late July; a single completion. No UKMTO or JMIC product addressing the island had been issued by this update's cut. The edition's readings, directions and access states are unchanged; they are re-read against the full month at the October edition.

Basis. Evidence after publication (Methodology v1.0: later documents and data generate a dated update, never a silent edit). Sources: Reuters (Aden/Dubai dateline, 11 September), Saudi energy ministry statements via wires (CNBC, CNN, Reuters), Kpler counts via Reuters (labelled preliminary and revisable).

As published. The 11 September log entry read “Houthi forces captured Mocha on 10 September, hours after publication” and “at dawn on 10 September — after publication —”, and gave “Events after publication” as its basis.

Corrected. The capture occurred at dawn on 10 September, about 02:30–03:00 UTC (Yemen is UTC+3), roughly three hours before the September edition's release at 05:38 UTC. The first public evidence — preliminary intraday ship-tracking (from about 11:14 UTC) and the first wire reports (afternoon UTC) — appeared only after the release. Under Methodology v1.0 the update classification turns on the evidence available at publication, so the dated update stands; the entry's sequence wording and basis line were wrong. The methodology's stated cadence (the 10th at 09:00 UTC) and the actual 05:38 UTC release both post-date the capture, so the error does not depend on which timestamp is used.

Basis. WRI wording defect in the log entry itself: event time and evidence time were conflated. Verified against the production release timestamp (2026-09-10 05:38:31 UTC) and the dawn timing carried in Reuters reporting; flagged by an external review on 12 September and corrected the same day.

Before. The September edition's Suez / Red Sea assessment closed with a data cut before 10 September; its access-uniformity criteria named Israel-, US- and UK-affiliated vessels, and its outlook recorded no attack on Red Sea shipping since 24 August.

Updated. Update, not correction: at dawn on 10 September — after publication — Houthi forces captured the Red Sea port city of Mocha and were reported advancing toward the Hanish Islands (Yemeni government military sources via Reuters; acknowledged by both belligerents and cited by the UN envoy at the Security Council). The Houthi-run Humanitarian Operations Coordination Center stated navigation remains safe “except for Saudi vessels” — an attributed access condition adding a fourth named class to the identity criteria. Preliminary same-day ship-tracking showed Bab el-Mandeb crossings falling sharply on 10 September (partial-day, revisable); JMIC Update 095 records 112 transits over 6–9 September, all pre-capture. No closure of the strait is reported. The edition's readings, directions and access states are unchanged; they are re-read against the full month at the October edition.

Basis. Events after publication (Methodology v1.0: later documents and data generate a dated update, never a silent edit). Sources: Reuters (multi-version wire, corroborated by AFP, AP, BBC, CNN, Al Jazeera, The National), JMIC Update 095 (10 September), Kpler preliminary counts via press (labelled).

Update, 12 September 2026. This entry's publication-sequence wording (“hours after publication”) is corrected by the 12 September log entry: the capture preceded the edition's 05:38 UTC release by about three hours; what followed publication was the evidence. The update classification is unchanged.

Before. The August edition printed the cost dimension under its research label 'Priority Premium' in its key figures and direction rows.

Updated. Both released editions now print the dimension as 'Cost of secured passage', the name the methodology defines. Readings, direction calls, reasons and advisory counts are unchanged.

Basis. Vocabulary alignment with the published methodology (§4: re-labelling under a new version is not a regime change); owner-approved 5 September, applied with the 10 September methodology publication.

Before. The Panama Canal announced-measures figure stated that the maximum Neopanamax draft would step to 14.48 m (47.5 ft) from 1 October, per advisory A-29-2026 — accurate when written.

Updated. Advisory A-33-2026 (4 September) postpones that step until further notice, citing Gatun Lake levels and weather projections; the 14.63 m (48.0 ft) maximum remains in force. The figure now says so.

Basis. Superseding ACP advisory issued after the edition's last update (an update, not an error at publication). The September edition will carry the current schedule in full.

As published. Five Strait of Hormuz direction rows (cost of access, waiting time, allocation rules, access uniformity, recovery flexibility) printed context sentences — 'there is no priority-access auction', 'no formal queue metric is published' and similar — alongside deteriorating readings, rather than the bases those readings rest on.

Corrected. Each row now states its evidenced basis from the edition's research record: the war-risk ladder to 7.5-10% of hull with reported Iranian fees of US$150k-2M; loitering clusters and blockade interdictions; the narrowing of options after the 14 July blockade reimposition; NAVCENT/NCAGS security-coordinated allocation; and the identity-gated two-corridor pattern. All directions and the 6-of-6 count are unchanged.

Basis. Same WRI editorial defect corrected on the Suez cost row on 5 September: the page's one-line reason is the first sentence of the underlying research assessment, which in these rows was context rather than basis. The remaining rows have now been swept edition-wide, and the assembly tool warns when a non-stable row's reason leads with an absence statement. Flagged by an external review on 6 September; corrected the same day.

As published. The Suez / Red Sea 'Priority Premium' direction row showed the reason 'SCA does not operate slot auctions or formal priority pricing' alongside a deteriorating reading — a sentence describing why no canal-authority price exists, not why the reading deteriorated.

Corrected. The row now states the actual evidenced basis: the de facto access price for the corridor is war-risk insurance, press-reported surging from a ~0.3% baseline to 1-3% of hull value from late July 2026. The deteriorating reading and the 3-of-7 count are unchanged.

Basis. WRI editorial defect at publication: the page's one-line reason is the first sentence of the underlying research assessment, which in this case was context rather than basis. The full assessment always carried the war-risk basis with its citation. Flagged by an external review on 5 September; corrected the same day.

Before. The Panama Canal recovery-flexibility direction row still described a late-cancellation surcharge as applying from 1 September 2026, per advisory A-28-2026 — accurate when the edition was released on 23 August.

Updated. Advisory A-32-2026 (2 September) postponed that surcharge until further notice before it took effect. The narrative was updated on 3 September; the direction row now says the same. The deteriorating reading is unchanged (the Last-Minute service remains suspended and substitutions restricted).

Basis. Superseding ACP advisory (update, not an error at publication). Completes the 3 September A-32 update, which missed this row; flagged by an external review on 5 September.

As published. From the 3 September refresh until 4 September, the edition's direction strips showed Suez / Red Sea at 2 of 7 and the Strait of Hormuz at 5 of 6 applicable access dimensions deteriorating, and the per-dimension lists carried reasons from a later research pass.

Corrected. The published August readings are Suez / Red Sea 3 of 7 and Strait of Hormuz 6 of 6, as released on 23 August; restored, with the per-dimension reasons from the August research inputs.

Basis. WRI assembly error: the 3 September refresh read direction inputs from the September research files instead of the archived August ones. Found by the daily review on 4 September and corrected the same day; the assembly tool now takes the month's archived inputs explicitly.

Before. The edition stated, per A-28-2026, that a late-cancellation surcharge (tariff 1050.IBC9) applies from 1 September to cancellations seven days or less before transit.

Updated. ACP Advisory A-32-2026 (2 September), issued after this edition's data cut, postpones the surcharge until further notice; it was not applied on 1 September. A new implementation date will come by a future advisory.

Basis. ACP Advisory to Shipping A-32-2026 (2 September 2026); IMF PortWatch, 3 September vintage. Changes caused by documents and data issued after publication — recorded as an update, not a correction.

As published. The edition stated the record Panama Canal slot-auction premium as US$4.6M (reported 14 August). A higher premium of about US$5.3M was paid at an auction on 24 August — inside the edition's 26 August data cut — and had been reported before the 26 and 29 August refreshes.

Corrected. Record ≈US$5.3M (auction 24 August; press-reported; confirmed to press as the highest ever by the ACP's designated administrator on 28 August). No reading or access state changed.

Basis. Bloomberg, 25 August 2026 (syndicated by gCaptain and the South China Morning Post); Panamá América/EFE and AFP, 28 August 2026. The ACP publishes auction volumes and a registered-access bid distribution, not individual winning bids, so the figure is press-reported.

Update, 30 August 2026. ACP Advisory to Shipping A-31-2026 (issued 28 August, effective 30 August for booking dates from 13 September) is noted under announced measures as governing from the September edition; it post-dates the data cut and changes no August reading. The Suez/Red Sea narrative figure for Bab el-Mandeb (26.9/day) was aligned to the table (27.8/day, same PortWatch vintage).

Before. The comparable-core row for the Suez Canal was blank (no PortWatch data was returned at the 26 August assembly); direction counts included dimensions shown as not applicable (Suez 3 of 8); the Panama advisory count was labelled as a count of rule advisories.

Updated. Suez row filled and the Suez headline aligned to the same PortWatch vintage (29 August: 40.9 transits/day; the 26 August vintage read 40.3); direction counts now exclude not-applicable dimensions (Suez 3 of 7); the advisory count is labelled as an activity count from an automated title screen, with the counted list published as CSV; ACP advisories cited in the form the ACP uses (A-29-2026).

Basis. IMF PortWatch daily chokepoint data, 29 August 2026 vintage; ACP Advisories to Shipping. No published reading or access state changed.

As published. The edition as released on 23 August stated that the Neopanamax maximum draft would fall to 48.0 ft on 26 August and to 47.5 ft on 3 September, and omitted the 15 September reduction to 23 Panamax booking slots per day (34 to 32 in total).

Corrected. Draft steps corrected to 2 September and 1 October; the 15 September Panamax slot step (23 slots per day, 32 in total) added.

Basis. ACP Advisory to Shipping A-29-2026 (20 August 2026), which supersedes the dates in A-25-2026. The edition had cited the earlier advisory where the later one governed.

Update, 26 August 2026. The same refresh incorporated ACP Advisory to Shipping A-30-2026 (25 August 2026), issued after release: the Special Competition mechanism for cancelled slots is noted under Recovery flexibility, and the 2026 advisory count moved from 11 to 12. These reflect a document issued after publication and are not corrections.