At dawn tomorrow, exactly two years will have passed since Bayesian sank off Porticello, claiming the lives of seven of the 22 people on board — Mike Lynch, his daughter Hannah, Jonathan and Judy Bloomer, Christopher and Neda Morvillo, and chef Recaldo Thomas — and the investigation into the sinking is entering what could prove to be its decisive phase. The wreck has been recovered and remains under judicial seizure in the port of Termini Imerese, where it has finally been possible to examine it in its entirety. These inspections will have to provide an answer to the question that has remained open from the outset: what sequence of events turned the heeling of a 56-metre sailing yacht into a sinking? Today, however, that answer matters for much more than reconstructing the casualty itself. Establishing the respective roles played by the wind, the yacht’s configuration, the raised centreboard, water ingress and the handling of the emergency means determining where any criminal and civil liability may lie and, consequently, who could ultimately face convictions, compensation claims and recovery actions potentially worth hundreds of millions. In the background is also the approximately €456 million claim brought by The Italian Sea Group against the yacht-owning company and several members of the crew.

The Termini Imerese Public Prosecutor’s Office opened an investigation into shipwreck and multiple manslaughter on 19 August 2024, the day of the casualty. Captain James Cutfield, engineer Tim Parker Eaton and sailor Matthew Griffiths, who was on watch that night, remain under investigation, a status that does not in itself imply any finding of responsibility. During the initial phase, attention focused primarily on the yacht’s preparation for the approaching bad weather and the management of the emergency. Giovanni Costantino, then CEO of The Italian Sea Group — the company that acquired the Perini Navi brand in 2021 but did not build Bayesian, which was launched in 2008 — publicly maintained that the sinking resulted from a chain of operational errors. The picture, however, has progressively become more complex. In May 2025, the preliminary report issued by the UK’s Marine Accident Investigation Branch, which has jurisdiction because the yacht was British-flagged, introduced an element likely to weigh heavily on the reconstruction: in the configuration Bayesian was in on the night of the casualty — sails stowed, engine running and centreboard raised — its Angle of Vanishing Stability was calculated at 70.6 degrees and, according to the MAIB, a lateral gust exceeding approximately 63 knots could have heeled the yacht beyond the angle from which it could no longer right itself. More significant still is the preliminary report’s finding that this vulnerability was not apparent from the stability information available to the crew. The issue therefore extends beyond the conduct of those in command to what the captain and crew could reasonably have known about the yacht’s behaviour in that particular configuration.

The interpretation of the weather event has also changed. In the hours immediately after the sinking, reports referred to a waterspout and an exceptionally violent weather event. Preliminary findings from the meteorological assessment commissioned by prosecutors and reported in 2026 instead pointed to a squall, with a rapid increase in wind intensity, without necessarily identifying a phenomenon that would have made the loss of the yacht inevitable. The distinction is substantial: investigators need to establish whether Bayesian was overwhelmed by unforeseeable conditions or whether the sinking resulted from the interaction between severe weather, vessel configuration, stability characteristics and emergency management.

The recovery of Bayesian in June 2025 transformed the scope of the investigation. Until then, the yacht had been lying on her starboard side at a depth of 49 metres, preventing a complete examination of the portion of the hull resting on the seabed. Once raised to the surface, following the removal of the mast required for the lifting operation, investigators were able to move beyond AIS data, witness statements, documentation, meteorological information and underwater imagery to a direct physical examination of the hull and onboard systems. Among the elements at the centre of the latest investigations is the 60-tonne retractable centreboard, which allowed the yacht’s draught to vary from approximately four to almost ten metres and which is believed to have been raised at the time of the sinking. The question is not whether it was appropriate in principle to keep the centreboard raised while at anchor, but rather to quantify how much that configuration affected the yacht’s transverse stability when the wind exerted a strong heeling moment on the hull, superstructure and mast.

Above all, investigators still need to establish where the water entered. Before the wreck was recovered, an internal glass panel between the engine room and the control room had been found cracked and deformed towards the latter. This was considered consistent with greater pressure on the engine-room side and therefore with the possibility that the engine room flooded first. It does not, however, explain through which opening the water entered that space. This is probably the technical heart of the investigation: heeling alone does not explain the sinking. The final reconstruction may therefore identify not one single cause, but a chain of events.

Understanding why she sank means establishing who pays

It is against this technical reconstruction that the most economically significant aspect of the case takes shape. Bayesian carried the types of insurance normally associated with a large yacht: Hull & Machinery and Protection & Indemnity. The former essentially covers the physical asset — hull, machinery and equipment — and Bayesian’s policy is reported to have had a limit of approximately €30 million. P&I, by contrast, covers the owner’s liabilities towards third parties and may include death and personal injury claims, legal expenses, environmental liabilities and wreck removal; in Bayesian’s case, this cover has been reported as being provided by British Marine. It is this second area that could involve considerably larger sums: potential exposure of around €400 million has been reported, alongside insurance limits of up to €2 billion. Those figures need to be interpreted correctly, as a policy limit represents neither the value of the yacht nor an amount that will necessarily be paid, but simply the maximum contractual cover available. The recovery of the wreck, whose cost has been put at approximately €25–30 million, also illustrates the overall economic scale of the casualty.

The technical reconstruction therefore becomes crucial from an insurance perspective as well. Should the investigation conclude that errors in the operation of the yacht or management of the emergency were the principal factors, the centre of gravity of responsibility would remain within the operational sphere of the owner and crew. If, on the other hand, design issues, shortcomings in the stability documentation or responsibilities attributable to other parties were to emerge, insurers could consider recovery actions against those deemed liable. A different scenario again would arise if the sinking were shown to have been unavoidable because of an exceptional weather event. Adding another layer of complexity is the Convention on Limitation of Liability for Maritime Claims, which under certain conditions allows a shipowner to limit exposure for certain categories of claims. Whether it can be applied to Bayesian will, however, depend on jurisdiction, the nature of the individual claims and the conduct ultimately established by the investigations.

A separate legal dispute was added to this already complex picture in January 2026. The Italian Sea Group brought proceedings against Revtom Limited, the owner of Bayesian linked to Angela Bacares, and several members of the crew. The claim, initially reported at £400 million and subsequently quantified at approximately €456 million, is based on the contention that the sinking resulted from the operational management of the yacht and caused TISG substantial economic and reputational damage, including lost business opportunities. It should be remembered that TISG did not build Bayesian: the yacht was built by Perini Navi and launched in 2008, thirteen years before the group acquired the brand and other assets from the company’s insolvency proceedings. Last July, Azimut Benetti chairwoman Giovanna Vitelli described the initiative as “unheard of, unprecedented”, underlining how exceptional a damages claim of this scale against a client is within the yacht industry.

A few days later, however, TISG’s corporate situation changed radically. Giovanni Costantino, founder of the group and promoter of the legal action, stepped down from the operational leadership of the Carrara-based holding company, while its severe financial crisis led to the launch of a process aimed at bringing in new investors and potentially transferring control of the company. This introduces another uncertainty into the Bayesian litigation: it will ultimately be for the company’s future ownership structure to decide whether to continue the action initiated by Costantino and, if so, whether to maintain its existing grounds and the amount of damages sought.

The €456 million sought by TISG should not be confused with the P&I exposure. On one side are the insurance consequences of the sinking and compensation claims by the victims and other affected parties; on the other is a company claiming that it suffered its own economic and reputational damage. Here too, however, everything ultimately leads back to the same technical investigations: a finding that operational failures were the predominant cause could provide support for TISG’s case, whereas the emergence of possible design or documentation issues relating to the yacht would make the litigation considerably more complex.

What happens next

The Termini Imerese Public Prosecutor’s Office has until February 2027 to complete its preliminary investigation. The three crew members remain under investigation and, at present, there is no definitive judicial reconstruction of the casualty, nor has any criminal responsibility been established. The decisive step will be the filing of the technical expert reports, which will need to establish a coherent sequence linking wind intensity, the yacht’s stability with the centreboard raised, the information available to the crew, water ingress, flooding, blackout and the ability of those on board to react. Only then will prosecutors be in a position to decide whether to seek dismissal in respect of one or more of those under investigation, pursue criminal proceedings or order further investigations.

February 2027 is therefore an important deadline, but not necessarily the end of the case. Should the investigation lead to a trial, it would mark the beginning of a new phase, while the British technical investigation and the civil and insurance proceedings would continue in parallel. Two years after the sinking, the Bayesian case therefore revolves around a limited number of fundamental questions: how strong was the wind that actually struck the yacht, how much did the raised centreboard affect stability, what did the crew know about the yacht’s stability characteristics in that configuration, where did the water enter and how much time was available to react? Above all, investigators will have to establish which links in the sequence that led to the sinking were foreseeable and avoidable. The answers will determine not only any criminal liability, but also compensation for the victims, insurers’ recovery actions and the prospects of TISG’s claim. Two years after Porticello, understanding why Bayesian sank now also means establishing who will ultimately bear the financial cost of the casualty.

The conclusions of the Italian and British investigations are also likely to lead to a review of the rules and procedures governing the safety of large pleasure vessels, both in terms of construction and certification and in their operational management. As has often happened throughout the history of maritime safety, the analysis of what failed that night may therefore contribute to new rules intended to reduce the possibility of a similar chain of events occurring again.