A human process failure involving conflicting weather and current data was the dominant cause of the scoring error that disrupted the 2026 Newport Bermuda Race, according to the event organisers’ investigation.
The Bermuda Race Organising Committee (BROC), its Technical Committee and US Sailing Offshore Office have published their investigation into the erroneous Forecast Time Correction Factors (F-TCFs) issued to the fleet on the morning of the June 19 start.
The investigation identifies two separate causes — a race-morning procedural error and a previously unidentified defect in the Expedition routing software.
However, BROC says the human error was responsible for the “great majority” of the scoring impact, while the software problem made a much smaller contribution.
Conflicting Weather Files Left Active
The primary problem occurred when additional ECM and Mercator GRIB weather and current files were downloaded into Expedition for a planned comparison of forecasting models.
They were not intended to form part of the official F-TCF calculations.
However, the investigation found that the files remained active alongside the official HRRR, GFS and RTOFS data when approximately 140 boats were routed on race morning.
As a result, Expedition was simultaneously working from competing wind sources and competing current sources.
According to the report, this introduced an effectively random element into the routing calculations and consequently produced erroneous predicted elapsed times and F-TCFs across the fleet.
The error escaped detection because the resulting routes appeared plausible and sample F-TCF checks against certificate ratings did not raise sufficient concern.
A genuine ridge of high pressure affecting the course also provided what appeared to be a credible explanation for some unusual differences between boats.
Competitor Query Exposed Scoring Problem
The problem was not identified until June 23, four days after the start, when the owner/operator of two Oakcliff Farr 40s questioned substantially different F-TCFs assigned to boats expected to have similar performance.
Running the two boats individually produced markedly different factors from those originally published, triggering the wider investigation.
All 140-plus boats were subsequently rerouted individually using only the official forecast data, and corrected F-TCFs were delivered to race officials on June 24.
Following redress hearings, the International Jury directed that the results be published using the corrected factors. The investigation says those results have since been independently reproduced and verified.
Expedition Software Bug Also Discovered
The investigation separately uncovered a defect in Expedition's “fleet mode”, which allowed multiple boats to be routed simultaneously.
Expedition developer Nick White identified and corrected the coding problem and released an updated version of the software.
However, analysis found that the fleet-mode defect produced deviations roughly one-fifth the size of the combined race-morning error and affected considerably fewer boats.
BROC therefore considers it unlikely that the software defect materially changed the outcome of previous Newport Bermuda Races or other events using Expedition fleet-mode F-TCF scoring.
Bermuda Fallout — The Newport Bermuda Race organisers have published their investigation into the F-TCF scoring error and outlined safeguards planned for future editions. Photo: BRF/Stephen R Cloutier
New Safeguards Proposed
The report recommends a series of changes before future F-TCF calculations, including restarting Expedition immediately before official race-morning routing and introducing a written checklist specifying exactly which forecast files and software settings should be used.
It also proposes that F-TCFs be independently generated by two operators using separate Expedition installations and reconciled before being released.
Further measures include stronger automated checks, additional operator training, retained audit logs and sufficient time between calculating the factors and the race start to allow proper verification.
Future of F-TCF and ORR Under Review
Significantly, the investigation has prompted a wider examination of the scoring system itself.
BROC says it will consider whether the additional fairness claimed for forecast-based F-TCF scoring justifies its operational complexity and reliance on specialist software, forecast models and manual procedures.
The committee will also separately investigate competitors' concerns that the underlying ORR rating rule may favour slower or older boats over newer, faster designs.
BROC says this will require statistical analysis across rating bands and comparisons with alternative rating systems.
Despite the controversy, the committee maintains that neither of the identified causes represents a technical flaw in F-TCF methodology itself.
It says early analysis comparing F-TCF with fixed Performance Curve Scoring continues to indicate that forecast-based scoring is “materially fairer” for a race exposed to substantial variations in weather and Gulf Stream conditions.
The committee acknowledged responsibility for the incident, saying it “remain[s] fully accountable for the results of the race” and regretted the disruption and uncertainty caused to competitors.
Download the full report below as a PDF file
