Why Accidents Rarely Start with an Emergency | BITE

The emergency is often the last link in a chain of threats and errors recognisable in advance. Fourth article in our boat safety series: Threat and Error Management, explained through a concrete case.

Almost no accident starts with an emergency. It starts much earlier, with an unrecognised threat or a small error nobody corrects in time — and the emergency is only the last link in a chain that could have been broken earlier.

In the third article of this series we talked about making better decisions once a problem shows up. In this article we take a step back, to talk about what happens before a problem becomes visible: how to recognise in time the conditions that increase risk, and the errors that follow from them, before they produce a concrete effect.

A chain of steps, not an isolated event

The concept that helped us think in these terms is called Threat and Error Management (TEM), and it too comes from human factors research and civil aviation. The word that matters most here is probably Management: it's not enough to list risks — they need to be actively managed, in real time, as they arise.

TEM distinguishes three separate moments, and distinguishing them is exactly what helps you intervene before things get worse.

Threats are the external conditions that increase risk, even before anyone makes a mistake: strong wind, traffic, shallow water, crew fatigue, pressure to arrive at all costs by a certain time. They're often predictable — weather can be checked in advance, traffic can be estimated, fatigue builds up in observable ways — and precisely because of that, recognising them before they appear is already a form of management.

Errors are the actions or omissions that arise from those threats, almost always unintentionally: a wrongly set course, a check skipped because you were in a hurry, an instruction understood differently from how it was given. An error, on its own, often doesn't yet cause harm: what happens next is what counts.

The undesired state is the moment when the error starts producing a concrete, visible effect: the boat drifting towards rocks, the distance from another vessel shrinking too much, a manoeuvre started too late, the crew no longer sure what the priority is and starting to get confused. This is the point where, without intervention, the situation can escalate into a real emergency.

Why the sequence matters

The reason distinguishing these three moments is useful is that each offers a different opportunity to intervene, and the further upstream you act, the less effort it takes and the more margin you have.

Recognising a threat in advance — for instance, knowing a stretch of the route will have heavy traffic at a certain time — lets you plan to avoid or reduce it, before there's even a chance to make an error related to that threat. Recognising an error just made — a skipped check someone notices — lets you correct it before it has consequences. Recognising an undesired state as it forms — drift beginning, distance shrinking — still lets you recover margin, even though the room to manoeuvre is already narrower.

Waiting for the emergency to react means having let all these earlier opportunities pass. This isn't a judgement on anyone who ends up in an emergency — it happens, even to well-prepared crews — but an invitation to look further back along the chain, not just at its final link.

Asking the right questions before a demanding leg

Before a demanding leg like the ones awaiting us on the way to Lisbon — a stretch of open water, a passage with heavy traffic or a strong current — we ask ourselves a few simple questions, which are our practical application of TEM: what threats do we expect for this specific stretch? What's the most likely error, given that we're tired, or the crew has recently changed, or time is limited? And above all: how will we notice it in time, before it becomes an undesired state?

That third question is probably the most important, and also the easiest to overlook. Anticipating a possible threat or error doesn't help much if you don't then know what to watch for to notice it taking shape.

Conclusion: managing, not just identifying

The central point of this article, the most concrete and accessible of the whole series, is that emergencies rarely arrive out of nowhere: they're almost always the last link in a chain that included threats recognisable in advance and errors that could have been corrected before producing an effect. The management part — the "M" in TEM — is what really makes the difference: it's not enough to identify a hazard; you need to anticipate it, notice it as it forms, correct it, and if necessary recover margin once the situation has already partly deteriorated.

In the next and final article of this series we'll talk about communication — not so much distress calls, but how information about a recognised threat, a mistake made, or a developing situation first needs to flow well inside the boat, among the people on board.