EAMAC Concours Ingénieur common mistakes
EAMAC Concours Ingénieur · Understand the traps and how to avoid them.
In a French dissertation or résumé de texte, the candidate starts writing after only a quick read, then discovers mid-page that the question asked for a comparison or a critical discussion rather than a description, and has to patch the plan in the margin.
Under time pressure the reflex is to fill the page immediately; the concours format (often 3–4 h for one essay) feels generous, so candidates underestimate the cost of a wrong plan and skip the 10–15 min of reformulating the sujet and listing its présupposés.
What to do instead
Impose a fixed pre-writing ritual: 10 min to rewrite the question in your own words, list the key terms and their ambiguities, and draft a two- or three-part plan on a separate sheet before the first line of the copy. Check the plan against the exact wording of the sujet once, then never return to it.
In mathematics, the candidate writes the answer to a question without stating the hypotheses that make the theorem applicable - for instance applying the intermediate value theorem or a convergence test without checking continuity, monotonicity or the sign of the general term - and loses points even when the numerical result is right.
Concours problems are long and the candidate optimises for finishing; school habits reward the final value, so hypothesis-checking feels like padding. Also, several classical theorems are memorised as formulas rather than as implications with conditions.
What to do instead
Adopt the habit of writing the theorem name and its hypotheses on the line before using it ('f continue sur [a,b], f(a)f(b)<0, donc…'). When reviewing past papers, mark every point lost for a missing hypothesis and keep a one-page list of the theorems whose conditions are most often forgotten (IVT, Rolle, dominated convergence, comparison of series, Cauchy–Schwarz equality case).
In physics, the candidate plugs numbers into a formula before checking units and orders of magnitude, and produces an answer such as a current of 10^6 A or a photon energy in joules where an electronvolt was expected, without noticing.
The pressure of the written paper pushes towards speed, and calculators hide dimensional nonsense. Candidates also mix SI prefixes (mA vs A, nm vs m) and forget that a dimensional check is a free verification.
What to do instead
Before substituting values, do a literal calculation to the end, then a dimensional analysis of the final expression, then a numerical estimate with powers of ten. Write units at every intermediate step. Train on 5–10 exercises per week where the only task is to predict the exponent of the result before computing it.
In the English paper, the candidate treats the reading comprehension as a vocabulary test: they translate word by word, answer with full sentences copied from the text, and in the written expression produce a memorised essay on a generic topic instead of answering the specific question asked.
French schooling often trains translation and grammar drills more than skimming for argument structure; candidates also fear making grammatical errors, so they copy the text to stay safe, which is precisely what the marking scheme penalises.
What to do instead
Practise the three-step method: read the questions first, skim the text for the relevant paragraphs, then answer in your own words within the word limit. For the essay, spend 5 min on a plan with a thesis and two arguments tied to the prompt. Build a personal list of 30 linking words and 20 topic-specific terms (aviation, environment, technology) rather than generic essay templates.
In the informatique/logique section, the candidate writes code or a truth table directly on the answer sheet without tracing it on a small example, and submits an algorithm that fails on the empty list, the single-element list or a negative input.
Written exams reward visible production; candidates equate 'having written a loop' with 'having solved the problem'. Edge cases are invisible unless deliberately tested, and pseudocode conventions vary between textbooks.
What to do instead
Before writing, state the specification (input domain, output, invariants). After writing, hand-execute the algorithm on three cases: empty/degenerate, typical, and boundary. For logic puzzles, draw the truth table or a small diagram rather than reasoning in your head. Keep a checklist of edge cases taped to your desk during revision.
In sciences de l'ingénieur, the candidate answers a functional-analysis question with a component list ('there is a servo-motor, a gearbox, a sensor') instead of identifying the function, the chain of energy and information, and the associated performance criterion.
Candidates revise the technology (names of actuators, bus protocols) because it is concrete and easy to memorise, but the concours expects a systems approach: analyse the need, the functions, the flows and the requirements. The vocabulary of functional analysis (bête à cornes, pieuvre, SADT) is known but not actually used as a reasoning tool.
What to do instead
For every system studied, produce a one-page analysis: need, main function, energy/information flows, actuators and sensors with their role in the chain, and two or three performance criteria with their order of magnitude. Practise on real aircraft systems (flight controls, landing gear, fuel system) and always link each component back to a function.
In aeronautical technology questions, the candidate gives a definition without the associated order of magnitude or operational consequence - for example describing the lift formula but not knowing that induced drag decreases with span, or confusing indicated and true airspeed when discussing a climb performance question.
Revision is often done from glossaries, which give definitions but not the quantitative relations between them. Candidates also study aircraft systems in isolation from the physics that explains why they exist.
What to do instead
Build a 'numbers and relations' sheet per topic: lift and drag equations with the role of each term, standard atmosphere values at sea level and at 10 km, typical cruise Mach and altitude, bypass ratio effects. For each system, write one sentence of the form 'this exists because otherwise …'. Test yourself by explaining a phenomenon aloud in 90 seconds without notes.
In probability and statistics, the candidate confuses P(A|B) with P(B|A) or treats events as independent without justification, and in the algebra section applies a result valid only for diagonalisable or symmetric matrices to a general matrix.
These are the classic traps of introductory courses, and the concours deliberately tests them. Under stress, candidates pattern-match on familiar wording ('sachant que' is sometimes missed) and reuse a memorised theorem without checking its scope.
What to do instead
Systematically rewrite conditional-probability statements as a tree or a two-way table before computing, and state independence explicitly when assumed. In algebra, keep a table of theorems with their exact hypotheses (spectral theorem for symmetric real matrices, Cayley–Hamilton always, diagonalisability criteria). Do timed 20-question drills on these specific confusions.
At the oral interview, the candidate recites a prepared presentation about their background for five minutes, cannot explain why they chose this school and this profession beyond 'passion for aviation', and has no concrete answer when asked about a technical topic from their own file.
The oral is often prepared last, and candidates rehearse a monologue instead of anticipating questions. Motivation is treated as an emotional statement rather than as an argument backed by experience, and the file (TIPE, projects, internships) is not re-read before the interview.
What to do instead
Prepare a 90-second pitch with three concrete anchors (a project, an internship or a personal achievement, and a precise reason tied to the school's curriculum). Re-read your own file and prepare five questions you would ask about it if you were the examiner. Rehearse with a timer and a listener who interrupts you, and prepare two questions to ask the jury.
Across the written papers, the candidate spends disproportionate time on one difficult question, runs out of time for the easy final exercise, and leaves whole pages blank or answers without a final sentence of conclusion.
Concours subjects are designed with a difficulty gradient, and the first questions of a long problem feel like the place to prove oneself. Candidates also fail to time themselves because they practise by subject rather than under exam conditions.
What to do instead
Do at least three full past papers under strict timing before the exam, with a written rule: if a question resists for more than 8–10 min, write the result as admitted and move on. Reserve the last 10 min for checking units, hypotheses and the presence of a conclusion in each part. Note on the draft the time allotted to each exercise before starting.