Concours Geipi Polytech common mistakes
Concours Geipi Polytech · Understand the traps and how to avoid them.
In a physics exercise on kinematics, applying a constant-acceleration formula (v = v₀ + at, x = ½at²) to a motion whose acceleration is explicitly given as varying with time or position, instead of integrating a(t) or using energy considerations.
The uniform-acceleration formulas are the most heavily drilled in terminale, so candidates pattern-match on the word 'acceleration' and skip checking whether a is actually constant. The Geipi Polytech multiple-choice format rewards speed, which reinforces the reflex.
What to do instead
Before writing any equation, explicitly write on the draft 'a = constant? yes/no'. If no, set up the differential equation or use the work-energy theorem. Practise mixed sets where constant- and variable-acceleration items alternate so the check becomes automatic.
In chemistry, computing a pH or an equilibrium composition without first checking whether the acid is strong or weak, or whether the approximation 'weak acid, negligible dissociation' is valid, and reporting the approximate value as exact.
Candidates memorise the two simplified formulas (pH = −log C for strong acid, pH = ½(pKa − log C) for weak acid) and apply them by keyword rather than by verifying the underlying condition (dissociation rate below about 5 %, or C compared with Ka).
What to do instead
Make the validity check a written step: compute the dissociation rate or compare C/Ka, state the criterion, then use the simplified formula only if it passes. Otherwise solve the quadratic or use the exact charge balance. Include at least one exercise per session where the approximation fails.
In mathematics, treating a function's domain of definition and its variation table as independent steps: differentiating f, finding the sign of f′, and building the table without ever restricting the interval to where f is defined (denominators, square roots, logarithms).
The mechanics of differentiation are over-practised relative to the preliminary analysis, and the domain question is treated as a formality to be answered in one line at the top rather than as a constraint on every later conclusion.
What to do instead
Adopt a fixed order: domain first, written as an interval or union of intervals; then derivative; then sign of f′ only on that domain; then limits at the boundary points of the domain, not at ±∞ by default. Mark boundary values on the table explicitly.
In mathematics, answering a probability question about a compound event by adding probabilities of the elementary events instead of multiplying along a branch of the tree, or by using P(A∪B) = P(A) + P(B) without checking incompatibility.
The addition rule is introduced first and is visually simpler; the multiplication rule requires constructing a model (tree, conditional probability) that takes an extra step under time pressure. The word 'or' in French ('ou') is mapped to addition and 'and' ('et') to multiplication without checking the actual structure of the experiment.
What to do instead
Force a tree diagram or a contingency table for every probability item before computing. Write the event in set notation, then apply the rule that matches the notation: union with the inclusion–exclusion formula unless disjointness is proved, intersection with the product rule only under independence or conditioning.
In physics, using the full distance travelled in a work-energy balance when the force is not applied over the whole path (for example friction acting only on part of the track), or omitting the work of a force that does act, such as the normal reaction doing no work but friction doing negative work.
The energy method is taught as a shortcut that avoids tracking forces step by step, so candidates write 'ΔEc = W' and then list forces from memory of the situation rather than from the actual path segments described in the statement.
What to do instead
Split the path into segments where the set of forces is constant, and write the energy balance separately for each segment with its own displacement. Draw the force diagram for each segment. Verify the sign of each work term against the direction of motion.
In STI2D/STL mechanics and energy teaching units, confusing the functional analysis of a system (which function is performed) with the structural/kinematic analysis (how it is performed), and therefore proposing a solution that does not answer the specification it was asked to satisfy.
The 'analyse et conception' approach is assessed through open questions where the expected answer is a justified choice against a cahier des charges, but candidates answer with a description of the existing solution or with a generic improvement, because the link between requirement and solution is rarely made explicit in class exercises.
What to do instead
For each open design question, restate the requirement from the specification with its criterion and level, then state the proposed solution, then justify it against that criterion. Use the three-column form (exigence / solution / justification) as a writing template until it becomes habitual.
In STI2D/STL, reading a characteristic curve or a data table without identifying the operating point imposed by the rest of the system, and quoting the maximum efficiency or maximum power value as the answer to a question about actual operation.
Efficiency and power curves peak at a single point that is easy to read and visually salient, whereas the operating point requires intersecting the load characteristic with the source characteristic, a step that is often skipped when the graph is presented alone.
What to do instead
Always identify both the source characteristic and the load characteristic before reading any value, and mark their intersection on the graph. State in one sentence why the operating point is not the maximum-efficiency point. Practise with items where the load line changes and the answer moves.
In STI2D/STL digital and electronic content, describing the behaviour of a combinational or sequential logic system by listing the states it passes through rather than by deriving its truth table or state table from the given expressions or schematic.
Simulation habits lead to a narrative 'when the input goes high, the output goes low' that can be right for one case and wrong for another; the systematic table is seen as redundant work when the answer seems obvious.
What to do instead
Write the truth table or state table first, with all input combinations in a fixed order, then read the answer from the table. For sequential systems, state the initial state explicitly and derive the next-state column before describing behaviour in words.
In STI2D/STL materials and structures, choosing a material or a cross-section from a single criterion (for example the lightest material, or the strongest) while ignoring the other constraints stated in the specification, such as cost, recyclability, corrosion resistance, or stiffness rather than strength.
Material selection is often taught through Ashby-type charts where one index dominates, and the specification's multiple criteria are read as context rather than as constraints that each eliminate candidates.
What to do instead
Convert every requirement in the specification into an explicit criterion with a direction (maximise or minimise) and, where possible, a numerical threshold. Evaluate the shortlisted candidates in a comparison table against all criteria, then justify the final choice including the trade-off made.
Across both scientific and technological papers, spending the same time on every question in a multiple-choice paper with negative or no partial credit, so that a long calculation early in the paper consumes time needed for several short items later, and guessed answers are left blank or answered randomly without eliminating options.
School assessments usually reward completing every question with full working, so candidates carry over the habit of finishing each item before moving on; the scoring rule of the competitive paper, which rewards correct answers and penalises neither blanks nor wrong answers in most items, is not internalised.
What to do instead
Before the paper, note the number of items and the time per item, and set a hard per-item ceiling. On a first pass, answer only items solvable in under the ceiling; mark the rest. On the second pass, eliminate options numerically or dimensionally before guessing, and never leave an item blank when there is no penalty for a wrong answer.