Investigators say Bayesian had an unrecognised vulnerability to high winds, never assessed at anchor with sails down, and that an unforecast downburst knocked her over in under 15 seconds.

There are reports that the yachting world reads out of duty and reports it reads with the door shut and the kettle forgotten, and the Marine Accident Investigation Branch's final word on Bayesian, published this morning as report 22/2026, is very firmly the second kind. More than two years after the sailing yacht foundered off Porticello on the north coast of Sicily, taking seven lives including the British technology entrepreneur Mike Lynch and his daughter Hannah, the investigators have set out, minute by minute and calculation by calculation, how a big, modern, well-crewed yacht lying quietly at anchor on a summer night can go from upright to lost in less time than it takes most of us to find a head torch.

The bare timeline is brutal enough. At 0406 on 19 August 2024 Bayesian was at anchor half a nautical mile south-east of Porticello when, in the MAIB's words, "a violent downburst of hurricane-force winds knocked the vessel onto its starboard side." Chief Inspector of Marine Accidents Rob Loder said that "in just 14 minutes, calm conditions had given way to winds of at least 68 knots, which in less than 15 seconds pushed the yacht onto its side." Unable to right herself, she flooded and sank in 50 metres of water. Six guests and nine crew escaped as she went down; six guests and one crew member did not. None of the liferafts inflated or released automatically, so the crew had to launch one by hand, and it was into that single raft that the fifteen survivors were gathered before a nearby vessel picked them up.

The heart of the report is not the weather, though, but the arithmetic of stability. The investigators found that, with no sails set, Bayesian's windage and stability characteristics left her vulnerable to a knockdown in severe winds, a vulnerability shaped by her loading, the apparent wind angle and the position of her centreboard, and one that "had not been identified in the originally approved stability information available to the crew." Put plainly, the crew were never told, because nobody had ever worked it out, that the one condition sailors instinctively regard as the safest of all, lying to an anchor with everything furled, was the one that could catch this particular boat out.

Nor did the forecast help them. The MAIB judged the forecast for the area "predominantly correct" but said it "did not indicate the risk of downbursts from the thunderstorms predicted", while the guidance published for mariners did not give enough detail on what such weather could do to support decisions on board. That conclusion sits awkwardly beside the line taken by prosecutors in Sicily, who have pointed at the crew; the British investigators, whose job is to prevent the next accident rather than apportion blame for the last, are clear that the people on board had no warning of the risk they were running.

The recommendations go to four bodies. The Red Ensign Group is asked to study the vulnerabilities the report identifies and amend its Yacht Code and guidance accordingly; the Maritime and Coastguard Agency is asked to add the risks of downbursts and severe gusts to the yacht master training syllabus; the Met Office is asked to raise with the World Meteorological Organization how low-likelihood, high-impact hazards should appear in weather data services; and the UK Hydrographic Office is asked to update its publications on extreme weather. The branch has also issued a safety flyer to the yachting industry and prepared a paper for the International Maritime Organization, and Mr Loder's plea was blunt: "I urge everyone with a part to play, including regulators, training providers and weather services, to read our report and act on it."

For those of us with rather smaller boats and rather fewer crew, the lessons scale down uncomfortably well. Few owners have ever seen a stability assessment for their boat at anchor, bare-poled, beam-on to a sudden gust; fewer still keep the hatches shut on a warm night in a quiet bay; and almost nobody, if we are honest, has ever wondered whether the liferaft cradle would actually let go if the boat went over. The Mediterranean summer thunderstorm, the one that blows through a crowded anchorage at three in the morning and is gone by breakfast, is not a rare beast, and the Bayesian report is a reminder that the forecast for a region and the weather in your particular bay can be two very different things.