Selection Simulator Sessions: What the Assessor Really Scores Behind the Flight Path

Poste d’instructeur de simulateur vu de dos, panneau éteint et voyants ambre dans une pénombre bleutée

The candidate leaves the session convinced he has failed: a one-dot deviation held too long on final, a heavy-handed flare, a go-around that had to be flown with both hands. He is selected. The candidate before him tracked the localizer to the needle and is never called back. There is nothing irrational about this. A selection simulator session is not a proficiency check: it is a sampling of behaviours over a very short window of time, and the flight path only serves as a medium to make those behaviours visible.

A Sampling of Behaviours, Not a Check

A proficiency check, or a practical skill test, rests on a published standard: the list of exercises is known in advance, the tolerances are written down, and the outcome is binary because it bears on privileges. A selection session has none of these properties. It grants nothing and withdraws nothing, and its result is not a pass/fail threshold but a rank: the candidate is compared against others in the same pool, against a grid the operator built to predict something that has not happened yet — how this person will behave after a type rating, on the line, in a crew they will not have chosen.

This difference in purpose governs everything else. An examiner is looking for compliance; a selection assessor is looking for a slope. He is not asking where the candidate currently stands, he is asking how fast that candidate moves when corrected. This is the source of a fact that many find disconcerting: a well-handled error is often worth more than a clean sequence. A clean sequence tells you nothing beyond itself; a handled error reveals the repair mechanism, and it is that mechanism the operator is buying, because it is the one that will matter later, on the line, when no one is looking over anyone’s shoulder anymore.

The observation setup itself makes this clear. The session is rarely scored by a single person: an instructor or type-rating examiner usually holds the technical grid, a line pilot from the company frequently occupies the right seat, and an observer from recruitment may follow the interaction without holding the technical grid. When this setup is in place, the written notes are cross-checked after the session, and a candidate who has only convinced one of these observers rarely passes.

Regulatory Check and Selection Simulator: Two Exercises Habit Tends to Confuse.
What Is at Stake Regulatory Proficiency Check Selection Simulator
Purpose Verify that already-held privileges are maintained Predict an aptitude to learn and to hold a seat
Standard Listed exercises, published tolerances, known in advance Behaviour grid, scenario not disclosed
Outcome Satisfactory or unsatisfactory, with effect on the licence Rank within a pool, with no effect on any licence
Status of Precision Explicit criterion, bounded by tolerances Indirect indicator of remaining spare capacity
Handling of Error To be corrected to stay within tolerances Raw material of the assessment, sometimes deliberately provoked
Aircraft Familiarity Type rating held and current Often zero hours on type, and the assessor knows it
What Makes the Difference Compliance with the standard The conduct adopted once the standard is no longer enough

The selection session isn’t about flying perfectly – it’s about showing how you handle mistakes and manage workload. The assessors care more about how you recover from errors than about small deviations in flight path.

The Grid Is Built from Verbs, Not Adjectives

The grids used today rest on a competency framework and on observable behaviours. The distinction sounds academic; it is decisive. An assessor is not allowed to write « good pilot », « at ease » or « lacks confidence »: these are impressions, they cannot be verified and cannot be defended before a third party. He must record what he saw or heard, with a verb and a moment — « announced the failure before acting », « did not request a time delay as the workload was rising », « read back the descent clearance without cross-checking it on the chart ». The consequence is blunt: whatever produced no observable produces no score, and a blank column never reads as a success.

The framework used in crew training breaks the job down into roughly ten competencies: application of knowledge, application of procedures and compliance with regulations, communication, flight path management through automation, manual flight path management, leadership and teamwork, problem solving and decision making, situation awareness and information management, and workload management. Count them: only two competencies concern the flight path, and one of the two deals with mode management rather than hand flying. All the remaining scoring surface is played out through speech, through the order in which actions are initiated, and through how the candidate manages his own time.

Each competency is backed by graded typical behaviours, and the same act is not worth the same thing depending on when it occurs: asking the other pilot to take the radio during a cruise checklist demonstrates nothing, asking for it in the seconds following a failure demonstrates an accurate read of one’s own workload. The grid is not a list of actions to produce, but a set of reference points against which conduct is situated in context.

On an A320 approach at 1,000 feet, a candidate notices a 1.5-dot deviation on the localizer. Instead of silently correcting, they say: “One dot left, correcting, we’ll confirm back within the band.” This single statement scores points for communication, flight path management, and teamwork – all in one action.

What the Flight Path Says, and What It Does Not

Precision is not ignored; it is read as an indirect indicator. A deviation interests the assessor only through three things: its magnitude relative to the margin still available, its trend, and what it costs in attention. A small but steady deviation that never closes says it is not being seen. A clear deviation, seen, called out and closed with a single correction says exactly the opposite. And a correction that oscillates says something else again: the pilot is acting on the instantaneous value instead of the trend, which is an analytical flaw rather than a handling flaw.

Magnitude of a Deviation and Trend of a DeviationTwo behaviours on final approach. The solid curve clearly exits the tolerance band, then returns to it in a single correction and stabilises there. The dashed curve stays within the band for a long time but drifts without ever reversing, and exits late without ever having been called out. The second concerns the assessor more than the first.deviation0duration of final approachtolerance bandclear deviation, called out and correctedslow drift, never called out
It is not the magnitude that concerns an assessor, it is the trend. A clear deviation closed in one correction says it was seen; a gentle drift that never reverses says it is not being seen, and it exits the band at the worst possible moment.

The moment a deviation appears matters more than its size. If the flight path degrades precisely when the radio frequency becomes busy, the assessor does not record a handling flaw: he records a coupling between two tasks that should have stayed independent, in other words a workload management failure. He will note in the same way the disappearance of the scan on parameters during a reconfiguration, the mandatory call that gets dropped right when a decision is being made, or the switch to manual flying just before the busiest event of the sequence.

One fact remains that candidates underestimate: on a type they have never flown, manual handling is the one competency the assessor knows he is not measuring much of. He factors this in, and he weights it accordingly. Devoting the bulk of one’s preparation to manual handling therefore means working on the one column where the bar has been deliberately lowered.

Verbalisation Is the Only Raw Material

An assessor has no access to reasoning: whatever stays inside the candidate’s head does not exist for the grid. A silent pilot who understood everything and a silent pilot who saw nothing produce the same written record, and that record is blank. This is how technically solid candidates walk away with a mediocre assessment without understanding why: they were working off the record.

Verbalising does not mean commentating. Continuous commentary is a mirror-image flaw: it saturates the internal channel, prevents the other pilot from getting a word in, and is paid for immediately on the flight path. Useful verbalisation holds to three elements, rarely more: intent, decision, criterion. « We keep the autopilot in, descend to four thousand, and go around if the approach is not stabilised at the set height » contains all three. « Right, well, let’s see » contains none of them.

The approach briefing is the first place where this difference becomes visible, and it is scored before the session has really begun. Reciting the chart in the printed order demonstrates only reading ability. Naming what could go wrong that day, fixing in advance the points at which a decision will be taken, stating how tasks will be shared if that decision falls due: these are three observables produced cold, without workload, and they will still count once the workload arrives. A candidate who announced his go-around criterion before the approach turns a difficult decision into the simple execution of a commitment already made.

The Error Is Not the Fault: It Is the Loop That Counts

The scenario is written to produce errors, out of methodological necessity: without a deviation, half the competencies on the grid remain unobservable. If the planned sequence is not enough, the assessor adds workload — a late runway change, a congested frequency, a cabin call, a minor failure at the worst possible moment. Reading this escalation as a punishment is a misreading of the nature of the exercise.

What is scored is the loop: detection, statement, decision, correction, verification of the recovery. Each of these steps leaves a trace, and each can be missing independently of the others. An error detected but not stated deprives the crew of its second pair of eyes. An error stated but not resolved lets the situation drift during the discussion. An uncorroborated correction creates the false belief that the problem is solved when it is not. Conversely, an error that was never seen leaves the realm of behaviour and enters that of physics: it becomes an undesired aircraft state, whose recovery is no longer chosen but endured.

The Two Outcomes of an Error During a Selection SessionAn error appears. If it is detected, announced and resolved, it is corrected and then verified: each of these five steps is an observable behaviour the grid can score. If it goes unnoticed, it turns into an undesired state whose recovery is endured, and the grid column stays blank.threat, then deviationerrorseen, stated, resolvedcorrected, verifiedneither seen nor statedundesired statefive scored observablesrecovery endured, grid blank
The grid does not score the absence of error, it scores the loop that closes it. The upper branch produces five observable traces where the lower branch produces none; this is why a handled error is worth more than a clean, silent sequence.

The trap: correcting the deviation in silence. From the assessor’s seat, a silent correction is strictly indistinguishable from a deviation that closed itself, and it deprives the monitoring pilot of the one piece of information that would let him help. The candidate believes he did well; the grid, meanwhile, records an unproven detection, an absent communication, and non-existent teamwork on the very same event.

The right reflex: state the deviation, the action, and the criterion for recovery, in that order and in a single short sentence. « One dot left, correcting, we’ll confirm back within the band » fills three columns of the grid, brings the other pilot into the verification, and makes the correction verifiable instead of leaving it open to interpretation.

Saturation Is the Object of the Test, Not Its Failure

A selection scenario loads the candidate until something gives, because the shape of the degradation is far more informative than the threshold at which it occurs. No one is eliminated for reaching their ceiling: everyone reaches it. What is scored is the order in which tasks are shed as the margin closes, and above all whether that shedding is decided rather than endured.

Workload and Capacity Margin Over the Course of the SessionThe imposed workload rises in steps throughout the scenario. Without shedding, the solid curve reaches and then exceeds the capacity ceiling and the flight path gives way. With two timely, decided sheddings, the dashed curve marks two plateaus and stays below the ceiling until the end.workloadcourse of the sessioncapacity ceilingno sheddingtwo decided sheddingsradio delegatedholding requestedthe flight path gives way
Everyone reaches their ceiling; what is scored is the shape of the climb. The two plateaus on the dashed curve are decisions, not pauses: each one frees up time before the margin is used up.

Delegating the radio, requesting a delay, asking for extended vectors, entering a hold to finish a procedure, declining an approach one will not be able to prepare properly: all of these are positive actions on the grid, never admissions of weakness. They show a pilot who knows the cost of what he is accepting. The classic hierarchy — fly the aircraft, navigate, communicate — only interests the assessor if it is applied as an announced decision, not as a reflex recited at the debrief.

The mirror-image trap: accepting everything offered in order to appear cooperative. The shortcut, the late runway change, the expedited descent are not favours but additional workload, and the assessor often offers them precisely to see how far the compliance will go. A well-reasoned refusal, stated without aggression and paired with an alternative, scores better than an acceptance that degrades everything that follows.

The Debrief Weighs as Much as the Session

The ritual question — « how do you think your session went? » — is not a courtesy, it is a measurement. It assesses the gap between what the candidate perceived and what actually happened. The one who identifies the essential part of his own deviations will train fast: he already has the instrument for correction. The one who sees none of them will be costly, because he will first have to be taught to see. The one who invents deviations everywhere is no better off: self-criticism that no longer discriminates is worthless on the line.

The stance matters as much as the content: describing the facts without dressing them up or blaming the flight model is enough. A factual remark about the simulator’s fidelity, made once, is neutral and sometimes fair; repeated, it becomes a defence and closes the one channel through which the assessor could observe that the candidate is coachable. The final question — « what would you do differently? » — checks whether a different course of action was genuinely available at that moment, with the information available at the time.

The Same Session, Two Behaviours: What the Assessor Writes in the Column.
Moment Behaviour A Behaviour B Competency Affected
Approach briefing Chart recited in the printed order Day’s threats named, decision points fixed, roles assigned Leadership, situation awareness
Failure at acceleration altitude Immediate memory items, in silence Flight path stabilised first, failure announced, delay requested Prioritisation, communication
One-dot deviation on the localizer Corrected without a word Deviation named, correction announced, return within the band confirmed aloud Flight path management, communication
Late runway change Accepted without discussion to appear accommodating Declined, reason given, holding or extended vectors offered Decision making, workload management
Debrief question The machine did not respond like the aircraft Three deviations identified, including one the assessor had not noted Application of knowledge, self-criticism

Preparing What Is Actually Scored

The practical consequence is clear: the most cost-effective preparation is not preparation for handling. It concerns the mode logic of the flight deck being used, through which flight path management is read; standard calls and the division of roles between pilot flying and pilot monitoring, the language in which the grid is written; and a personal decision framework, whatever it may be, applied out loud rather than recited. Selection preparation courses work first on these verbal loops, for the same reason.

Two men in a cockpit discuss with flight instruments illuminated in the background.
Assessors evaluate performance beyond technical flight metrics.

The right reflex, upstream: set and write down your criteria before the session — what will trigger a go-around, what will cause an approach to be declined, what will cause the radio to be delegated. A criterion decided in cold blood is announced in one sentence at the briefing; the same criterion improvised under load costs several seconds of attention at the moment they are needed most.

A grey area remains. A session lasting a couple of hours is still a very short sample, on an unfamiliar aircraft, in front of assessors whose standardisation is real but imperfect. Operators know this: the session almost never decides alone, and is combined with an interview, group exercises and a review of the file. It is therefore worth checking, in the documentation the operator publishes and in feedback from past campaigns, what weight the session actually carries in the decision, who occupies the right seat, and whether that seat is held by a pilot briefed to execute exactly what is asked of him, or to help only if asked correctly.

For anyone preparing for a type rating in the months that follow, the exercise has a value that goes beyond hiring: the competencies examined in selection largely overlap with those the type rating will score next, and then those the periodic check will score throughout the career. The vocabulary of observable behaviours is learned only once, and it is never unlearned.

The applicable regulations, the operations manual, and the operator’s approved training documentation prevail over this article, which sets out only general principles of assessment; in the event of any discrepancy, it is the documentation of the operator concerned that governs.

Questions fréquentes

Should you know the type of simulator used before a selection session?

It helps, but it is secondary. The operator knows that a candidate not rated on the type has neither the reflexes nor the reference values, and weights the manual handling competency accordingly. Preparation time pays off far more on autopilot mode logic and on task-sharing, which are scored regardless of the type.

Does a poorly flown go-around eliminate a candidate?

Rarely by itself, if the decision was taken at the right time and announced. What eliminates a candidate is the absence of a decision: continuing an unstabilised approach, or waiting for the situation to decide for you. A rough manoeuvre decided early scores better than a clean manoeuvre initiated too late.

Should everything be verbalised during the session?

No, and continuous commentary is a flaw just as costly as silence. It saturates internal communication and is paid for on the flight path. The practical rule holds to three elements: intent, decision, recovery criterion — the rest can stay inside your head without any harm.

How should you answer when the assessor asks what you think of your own session?

By describing facts, in chronological order, including the unflattering ones. Naming three or four genuine deviations and stating what you would do differently, with the information you had at that moment, is worth more than an overall assessment. It is the precision of the self-criticism that is measured, not its severity.

Is the company pilot sitting in the right seat a help or a trap?

Neither: he is a measuring instrument. He generally executes what is clearly asked of him, and nothing that is merely assumed. How a candidate assigns him tasks, gives him a criterion, and verifies the result feeds directly into the leadership, communication and workload management competencies.

Can you decline a runway or a shortcut offered during the session?

Yes, and it is often the expected answer. These offers are frequently introduced to measure how far a candidate’s compliance will go. A well-reasoned refusal, stated calmly and paired with an alternative — a hold, extended vectors, the next approach — is scored as decision making, not as rigidity.

Commentaires

Leave a Reply

Découvrez l’écosystème FL330

FL330, c’est toute l’aviation professionnelle — explorez nos autres espaces.