Imagine, for a moment, that you are aboard a 56 metre luxury yacht anchored off Sicily. It is before dawn, there's some wind but the crew aren't worried, and you're coming up to take a look around. Suddenly, the yacht is knocked about as all the windows explode around you. Confused, startled, and scared, you run for a muster station, managing to get aboard a life raft moments before the yacht disappears beneath the waves. The whole event took longer to write up than it took to happen.

That's what happened to the survivors of the sailing yacht Bayesian this week, in an incident that has the sailing world alternating between grief, confusion, exasperation, and grasping at straws for an explanation.

The authorities and the experts they've called in for the search and investigation, naturally, are taking a careful and measured approach. So far, they've refused to state the cause of the incident – a correct call, since the cause is not yet known.

That hasn't stopped scores of amateur sailors, wanna-be naval architects, and even a few practising designers from climbing atop Mount Stupid to speculate about the cause of the wreck, spout incomplete and/or inaccurate information about ship stability, and to lay blame at the feet of the designer, the yard, the captain, the crew, and/or the now-deceased owner.

Lifting Keels and Stability

Much of the amateur speculation revolves around the fact that Bayesian had a lifting keel. To an armchair sailor or keyboard warrior, and even to some designers of much smaller craft, this is an obvious smoking gun: lifting the keel raises the centre of gravity, which reduces stability, which makes the boat less safe. There! Root cause found!

Well, no.

Let's go back to our stability basics.
(And, once you've read that, we'll continue here.)

A lifting keel's purpose is to allow a foil to be extended to a suitable span & chord that will create lateral lift to resist the leeway generated by sail forces when reaching or pointing.

Mounting ballast on the keel tip (point K) brings the centre of gravity (point G) down. This increases the righting arm (GZ) and therefore increases the sail-carrying power, making the boat faster while heeling less under sail.

Moving G down also increases the angle of vanishing stability, and shrinks the region of negative stability, making the boat harder to capsize and easier to right once capsized.

What the quick-to-speculate crowd forget is that all of this has been known to all engineers and naval architects in the field for centuries. The insurers and the classification societies know it too. It's not arcane secret knowledge. You cannot get Lloyd's, DNV, ABS, etc. paperwork for a lifting-keel yacht without stacks of carefully calculated stability curves that prove it is safe and stable with the keel up.

The slight increase in stability with the keel down is just a bonus that lets you carry more sail with a little less heel angle. And, in superyachts, it is *slight*. They sail at relatively low heel angles (can't be spilling the drinks, you know) and have computer-controlled winches that release the sheets at specific heel angles well below the heel angle that yields the maximum righting arm. And, as mentioned earlier, the lifting keel's main purpose is to be an efficient lifting foil to fight leeway. As you go to bigger and bigger boats, the contribution to overall stability that comes from form stability due to beam – i.e. the lateral shift of the centre of buoyancy B(Φ) when heeled to an angle Φ – scales up much faster than the contribution from the weight in the keel tip. Thus, form stability becomes hugely dominant over the relatively small vertical change in centre of gravity G that you get by extending or retracting the keel, when the boat is at low to moderate angles of heel.

You can't tell Perini Navi to go to Lloyd's with paperwork that says "Yeah, the yacht is totally fine with the keel down, but if it gets hit by a good gust when the water's only 5 metres deep it's gonna go right over and sink." Perini Navi will tell you to pound sand, and when you go to Lloyd's anyway, they too will laugh and stamp it with a big red "Rejected".

Setting the question of the retractable keel's effect on static intact stability aside – since, for this to be the root cause, several dozen of the world's best naval architects, surveyors, and construction managers would all have had to neglect centuries of well-known math and design codes, which seems unlikely – we are left with the usual five main causes of vessel loss. Fire, structural failure, downflooding, collision with other vessels, and collision with terrain. In this case, three of those can be ruled out with high certainty. Let us consider the other two.

Structural Failure

Yachts have been lost to structural failure before. The Cheeki Rafiki is the first that comes to mind, that tragic disaster having been caused by a worrying confluence of bad standards, bad design, bad construction, bad inspection, and bad maintenance.

Whether that might be the case here remains to be seen. Perini Navi's history as a builder suggests that it is unlikely, but it is never impossible.

Speculation about structural failure would be irresponsible until there is actual evidence in hand.

Downflooding

An excellent angle of vanishing stability doesn't mean much if the hull starts downflooding through openings before that angle is reached.

Take a look at any modern superyacht while it's at anchor and at play. How many gaping openings in the hull sides do you see? There are tender garages, swim platforms, private sea-level fold-out balconies. Any of these openings could easily admit water in tremendous quantities if submerged. This specific yacht does not appear to have had the enormous side openings common to some other recent designs, but its stability booklet - recently obtained by Peter Swanson at Loose Cannon - does list several ports and hatches that, if left unlatched, would cause critical flooding at heel angles in the 43° to 48° range – well within the boat's nominal region of positive stability, and also well within what a good hard gust on bare poles might be able to achieve.

There are, therefore, strict design and operating rules for such openings. The panels themselves are built like the hull or stronger, their perimeters are heavily reinforced, and heavy latches clamp them down on multiple gaskets whenever the wind and wave conditions might pose a risk. There are watertight doors between the compartments within the ship, which can be closed when conditions call for it. This boat's stability booklet, like those of all vessels with hull-side openings, is full of cautions to the master and crew about securing these openings and ensuring that the boat is not operated in a way that would risk downflooding.

Those of you who know your Hierarchy of Controls can see the problem. We've created a new hazard by design – elimination and substitution are gone. We've implemented engineering controls (latches, interlocks, position sensors, and the like) as much as possible. But they rely on an administrative control to work, and have to be overridden for the feature to serve its purpose. Some person has to watch the weather, watch the board, and flip the switches when necessary, and if that person ever neglects that duty, disaster is inevitable.

If the openings don't get closed in time, or are partly closed but not properly latched, you have a massive downflooding risk at heel angles where the keel isn't yet in play at all. And we are not talking splashes here; these are flood-the-compartment-in-ten-seconds quantities of water.

Once that water's on board, you have a free-surface-effect problem. Remember the MS Herald of Free Enterprise? That 7950-ton Ro-Ro forgot to close its bow door, waves started lapping over the car deck as it got up to speed, and all stability was lost within 90 seconds.

Nothing a designer or captain can do with ballast, pumps, or heroic efforts of crew can do a damn thing about a runaway free surface effect in a vessel that's started to downflood.

This may have been the case with the Bayesian. Material translated from Italian social media, purportedly written by senior members of its build team, seem to suggest that they believe crew negligence leading to downflooding is a likely cause. It is still too early to speculate further; not all the facts are known. We must wait for a proper engineering report to know for sure.