Bridge Resource Management

Ask any accident investigator what went wrong on the bridge and the answer is rarely "nobody knew the rule." It is almost always that someone had the information and it did not translate into the right action in time. Situational awareness is usually described in three levels, and it is worth holding onto the distinction because the exam will test which level failed, not just whether awareness existed at all.

Perception is simply noticing: the echo appeared on the radar, the depth alarm sounded, the VHF call was heard. Comprehension is turning that raw signal into meaning: that echo is a fishing vessel crossing from starboard, that alarm means the keel clearance has fallen below the margin set in the passage plan. Projection is running the situation forward: if nothing changes, where will both vessels be in six minutes, and does the passage plan still hold in twenty?

Investigations of collisions and groundings repeatedly find perception intact — the target was on the radar, the alarm did sound — while comprehension or projection failed. An officer sees a closing target and logs it without asking what it means for the next ten minutes; a depth alarm is silenced as a nuisance without asking why the margin has eroded. The practical lesson is that watchkeeping is not finished once something has been seen. It is finished once its meaning and its trajectory have been worked out and, if necessary, acted on.

Almost no serious incident has a single cause. What actually happens is a short sequence of small, individually survivable failures that line up — the "error chain." Take away any one link and the chain usually breaks before it reaches the accident.

• Ambiguity — an instruction, a chart note or a situation that could reasonably be read two ways, and nobody asks which one is meant.
• Distraction — attention pulled onto a secondary task (a phone call, paperwork, a minor alarm) while the primary task keeps running unmonitored.
• Fixation — locking onto one explanation or one task and filtering out evidence that contradicts it.
• Communication breakdown — an order given but not confirmed, or confirmed but not actually understood.
• Non-compliance — a deliberate or half-conscious departure from the passage plan, the checklist or the standing orders.
• An unchallenged deviation — someone notices the departure from the plan and says nothing, so it is silently accepted as the new normal.

The countermeasure the examiners are looking for is not "more vigilance" in the abstract. It is the specific habit of naming the link out loud: saying "I think we're fixating on the pilot boat and losing the ferry" is a concrete, examinable action, whereas "stay alert" is not. A bridge team with a shared vocabulary for these links can interrupt a chain at almost any point, because naming the failure turns a private unease into a team problem that someone is now responsible for resolving.

A bridge with a master, a pilot, an OOW and a helmsman has, in principle, several independent checks on any single mistake. In practice it often has only one functioning brain, because everyone junior to the person currently conning the ship assumes that person has already thought of what they are about to say. That assumption is the authority gradient, and it is dangerous precisely because it is invisible from inside it — nobody feels like they are staying silent out of deference; they feel like they are probably wrong.

Two procedural habits are the direct countermeasure. The first is closed-loop communication: an order is stated, repeated back in the receiver's own words, and then its execution is confirmed once it has actually happened. Silence after an order is not confirmation. The second is the two-challenge rule: if something looks wrong, it must be raised at least twice, in increasingly direct terms, before the junior officer is entitled — in fact obliged — to escalate or act independently. The rule exists precisely because a single polite question is easy for a busy or confident senior officer to miss or wave away.

What makes this examinable rather than just good manners is the time cost. Raising a concern, having it dismissed, raising it again and then escalating all take measurable minutes, and those minutes come straight out of whatever margin the situation had to begin with. Knowing the protocol is only half the answer — the other half is knowing how much of the available time it consumes, which is exactly what worked example three below asks you to calculate.

Fatigue does not announce itself. It erodes perception first — the level of situational awareness that feels most reliable — so a tired officer usually believes their judgement is intact right up until it demonstrably was not. Because self-assessment is unreliable, BRM manages fatigue with hard numbers rather than a feeling of tiredness, and the exam expects you to apply those numbers, not just recite them.

The two-part shape of that rule matters as much as the totals. It is entirely possible to log more than ten hours of rest in a day and still fail the requirement, because the rest was chopped into three or more short pieces by drills, cargo operations or admin — fragmented rest is worse for alertness than the same number of hours taken in one or two continuous blocks, which is exactly why the rule caps the number of periods rather than only the total. Worked example one below walks through a day that passes the hours test and fails the periods test, which is the version of this question the exam is most likely to ask.

The second half of workload management is distributing tasks before anyone reaches that fragile state, not after. A busy pilotage or a fog transit generates a fairly predictable rate of radar checks, position fixes and radio calls, and that rate can be totalled up and compared with what one person can sustainably deliver. When the total exceeds what a single watchkeeper can do without their standard dropping, the correct BRM response is to delegate and prioritise before performance degrades, not to have the OOW quietly try to do all of it and hope. Worked example two makes that comparison explicit.

Pilotage is one of the few situations where two people with real ship-handling authority are on the bridge at once, and BRM exists partly to keep that from becoming a source of confusion rather than redundancy. The pilot brings local knowledge — the set, the traffic, the berth — but does not take command of the vessel. The master remains in command throughout, which means the master remains responsible for the ship's safety even while largely deferring to the pilot's local expertise, and retains the authority and the duty to intervene if the pilot's actions appear to be putting the vessel at risk.

That relationship is formalised through the master–pilot exchange, held before pilotage begins and covering the intended passage plan, the ship's handling characteristics (loaded draught, manoeuvring speed, any known handling quirk), the tug and berthing arrangements, and any outstanding defect that could affect manoeuvring. Anything in that exchange that the master and pilot see differently has to be resolved there, on the record, before the vessel gets underway — not worked out informally halfway up the fairway when a course alteration is already in progress and neither party is fully sure what the other is assuming.

• Plan — the intended track, speed, and any critical decision points.
• Handling data — draught, trim, manoeuvring characteristics, minimum RPM for steerage.
• Tugs and berthing — number, position, and the intended mooring sequence.
• Defects and contingencies — anything not working as it should, and what happens if the plan cannot be followed.

The bridge team's job during pilotage does not stop once the exchange is complete. The OOW and master continue to independently monitor the vessel's position and the pilot's conning exactly as they would monitor each other, because the exchange establishes shared expectations, not a transfer of responsibility for checking them.

Modern bridges give a level of precision — GPS position to a few metres, an ARPA solution refreshed every few seconds — that earlier generations of watchkeepers never had, and that precision is genuinely valuable. The risk BRM tries to manage is not the equipment itself but the habit it quietly encourages: trusting a displayed answer because it has always been right before, rather than because it has been checked against something independent this time.

Automation complacency shows up in recognisable forms: accepting a GPS position without a periodic cross-check by an independent method, accepting an ARPA-derived CPA and course for a target without a visual or plot-based sanity check, or following an ECDIS route without noticing that a waypoint sits inside charted shallow water. In each case the system was not necessarily wrong — the failure was that nobody checked, because the system's normal reliability had quietly become an assumption of infallibility.

The examinable habit is simple to state and easy to skip under time pressure: before acting on an automated figure, ask what independent evidence — a visual bearing, a second fix method, a manual plot — would either confirm or contradict it, and get that confirmation before the figure is relied on for a decision with no room to recover from being wrong.

The three examples below use the numbers from the sections above — rest-hour arithmetic, workload against capacity, and the time cost of the two-challenge rule — to show how BRM questions that look like judgement calls actually reduce to a calculation once the situation is set out properly.