JEE Advanced common mistakes
JEE Advanced · Understand the traps and how to avoid them.
Substitutes straight into a memorised formula without first writing down the conditions under which it holds - e.g. writes R = u²sin2θ/g for a projectile launched up an inclined plane, or uses the thin-lens formula with the wrong sign for a virtual object.
Formulas are rehearsed as strings of symbols tied to a recognisable diagram, so the surface-level resemblance ('projectile', 'lens') triggers the equation before the assumptions (horizontal ground, paraxial rays, fixed sign convention) are checked.
What to do instead
Before any substitution, write one line listing the validity conditions and the sign convention you are adopting. If a condition fails, re-derive from the two basic equations (e.g. resolve along and perpendicular to the incline) rather than adapting a memorised result.
Writes constraint relations for connected bodies (pulleys, wedges, blocks on inclines) by visual inspection, then gets a relative-acceleration or tension sign wrong, producing a plausible but inverted answer.
The constraint is inferred from how the string 'looks' rather than from geometry, and the chosen positive direction differs between the two bodies, so one term silently flips sign.
What to do instead
Fix a coordinate axis on the diagram, write each body's position coordinate, then differentiate the length constraint twice. Check the result in a limiting case (mass → 0 or → ∞) before solving the equations.
Produces a numerical answer with the wrong power of ten or wrong unit because centimetres, grams, eV or atmospheres were fed into an SI formula, or an intermediate was rounded too early.
Under time pressure the unit conversion is done mentally and never written down, and the final number 'looks reasonable' so it is not sanity-checked against the physical scale of the system.
What to do instead
Convert every quantity to SI in a written first line, keep one extra significant figure through intermediates, and finish by estimating the answer's order of magnitude independently (e.g. a 1 kg block cannot accelerate at 10⁴ m/s² under a 1 N force).
Treats the equilibrium constant K as shifting with pressure, concentration or added catalyst, or answers a ΔG question using ΔG° (and vice versa) without distinguishing standard-state from actual conditions.
Le Chatelier's principle is remembered as 'the system opposes the change', and this gets misapplied to K itself rather than to the composition; ΔG and ΔG° are stored as one symbol.
What to do instead
Anchor two facts: K depends only on temperature (catalysts and pressure change the approach to equilibrium, not K), and ΔG = ΔG° + RT ln Q. When a question mentions K changing, ask explicitly 'what changed - T only?'
Predicts an organic product from the reagent name alone, ignoring the substrate's substitution pattern, the regiochemistry/stereochemistry demanded, and the stated conditions (excess reagent, temperature, solvent, aqueous vs alcoholic KOH).
Reaction-name recall is trained as reagent → product pairs, so the substrate is treated as a generic alkyl halide or alkene and competing pathways (SN1 vs SN2, E1 vs E2, ortho/para vs meta) are never compared.
What to do instead
For every organic question, write the substrate's structure with its key features (degree of substitution, leaving group, existing substituents and their directing nature) before choosing a mechanism. Then pick the major product by explicit comparison of competing pathways under the given conditions.
Balances atoms but not charge in ionic/redox equations, or ignores the limiting reagent and the acidic/basic medium when adding H₂O, H⁺ or OH⁻.
Balancing is done by counting atoms visually on each side; charge and medium feel like bookkeeping details, and limiting-reagent analysis is skipped when the numbers 'divide nicely'.
What to do instead
Balance redox by half-reactions: balance atoms other than O and H, then O with H₂O, then H with H⁺ (or OH⁻ in basic medium), then charge with electrons, and finally verify total charge on both sides is equal. Separately, convert every given amount to moles and divide by the stoichiometric coefficient before deciding the limiting reagent.
Squares both sides, clears a denominator, or takes a logarithm without checking the resulting roots against the original domain, and reports extraneous solutions (or discards valid ones) in an integer-answer question.
The algebra is treated as a sequence of legal operations, so the direction of implication (squaring is not reversible) is never tracked, and excluded values such as x > 0 for logarithms are not recorded at the start.
What to do instead
Write the domain/excluded set before solving, and mark every non-reversible step (squaring, multiplying by an expression, applying a non-injective function). Substitute each candidate root back into the original equation, or verify it satisfies the recorded domain, before finalising the answer.
Introduces a parameterised form - slope m for a line, (at², 2at) for a parabola, a general point on a circle - and silently loses the cases the parameter cannot represent (vertical line, m infinite, degenerate conic).
The parameterisation is chosen because it makes the algebra quick, and its coverage gap is invisible: no term in the working ever flags the missing configuration.
What to do instead
After choosing a parameterisation, state in writing which configurations it excludes and treat those as separate cases (e.g. check x = c for a vertical line, check D = 0 and the a = 0 degenerate conic). Verify the final answer set includes the excluded case if it satisfies the condition.
Finds f′(x) = 0, calls those points the maxima and minima, and reports them as the answer to 'range of f on [a, b]' or 'number of real roots', ignoring endpoints, non-differentiable points and monotonicity behaviour.
The derivative test is over-learned as a complete procedure, so the closed-interval hypothesis and the behaviour outside critical points are never part of the mental checklist.
What to do instead
Use a fixed routine: domain → endpoints → critical points (including where f′ does not exist) → sign chart of f′ → sketch. Compute f at endpoints explicitly and compare with the critical-point values before writing the range or counting roots.
Miscounts in combinatorics or probability by mixing ordered and unordered selections, double-counting cases already covered, or assuming outcomes are equally likely when they are not (e.g. treating the sum of two dice as uniform).
The counting is done by intuition on small cases and the symmetry assumption is never tested; identical objects and 'at least one' conditions are handled by ad hoc addition of overlapping cases.
What to do instead
Declare at first whether order matters and whether objects are distinct, and match that to nPr/nCr or to a labelled-box model. For 'at least one', use the complement. Verify by enumerating the same count a second way (or by computing a small analogue) before committing to the answer.