Approximation Techniques for JEE
On this page
Direct answer
When JEE Physics options sit at 4, 40, 400 and 4000, the paper is testing approximation, not long division. The core tools are the binomial truncation (1 + x)^n ≈ 1 + nx valid for x much smaller than 1, the small-angle results sin θ ≈ θ and tan θ ≈ θ with θ in radians, cos θ ≈ 1 − θ²/2 when the leading term cancels, and e^x ≈ 1 + x for small x. Alongside these sit two habits: converting percentage changes through differentials (a small change in l changes T as ΔT/T = Δl/2l), and sanity-checking any derived expression by pushing a variable to zero or infinity to see whether the answer collapses to a known result. Every JEE paper is written to be survivable without a calculator precisely because these moves exist.
What you must remember
- Binomial: (1 + x)^n ≈ 1 + nx for |x| << 1, any real n; so √26 = 5(1 + 1/25)^0.5 ≈ 5(1 + 0.01) = 5.05, and (1 − x)^n ≈ 1 − nx.
- Small angles (radians only): sin θ ≈ θ, tan θ ≈ θ, cos θ ≈ 1 − θ²/2; at 10 degrees sin θ is already within 0.5 percent of θ, which is why the pendulum formula survives small swings.
- Exponential and reciprocal: e^x ≈ 1 + x; 1/(1 + x) ≈ 1 − x; ln(1 + x) ≈ x — the trio behind RC and LR circuit estimates.
- Percentage change via logarithmic differentiation: for T = 2π√(l/g), a 2 percent rise in l raises T by 1 percent; for g at small height, g′ = g(1 − 2h/R), so each percent of R climbed costs 2 percent of g.
- Limiting-case audit: set m2 → ∞ in a collision formula and you must recover reflection from a wall (v′ = −v); if you do not, the algebra is wrong before the arithmetic is.
- Order of magnitude: keep numbers as powers of ten and fold coefficients at the end; most wrong option-eliminations fail on exponents, not coefficients.
- Second-order rescue: when a leading term cancels (two nearly equal forces, cos θ differences), keep θ²/2 terms — dropping them leaves zero and no answer.
One problem, three approximations
A satellite orbits at height h = 320 km above Earth (R = 6400 km). Exact percentage change in g would demand division by 6720; the binomial does it mentally: g′ = g(1 + h/R)^(−2) ≈ g(1 − 2h/R) = g(1 − 2 × 0.05) = 0.9 g, a 10 percent drop. Now ask for the orbital speed change: v = √(g′R′) ≈ √(0.9 g) × √(1.05 R), so v′/v ≈ √0.9 × (1.05)^0.5 ≈ 0.949 × 1.025 ≈ 0.97, a 3 percent fall — two chained binomials, no calculator.
The same discipline polices derived results. For a simple pendulum released from a small angle, the equation of motion becomes (g/l)θ = −θ̈ only after sin θ → θ; a student who keeps sin θ cannot integrate, and one who approximates cos θ ≈ 1 in the energy equation while hunting the θ² restoring term gets zero instead of the right answer. Approximation is not sloppiness — it is knowing which term survives at the order you are working, which is precisely what JEE Advanced tests when it asks for a fractional change rather than a value.
Where the tricks bite back
JEE Main usually rewards the first-order move directly — evaluate (0.998)^50 or find the percentage change in resistance when length stretches 0.5 percent. JEE Advanced punishes blind use: in a problem where two nearly equal magnetic fields oppose, the answer lives entirely in the second-order difference, and (1 + x)^n ≈ 1 + nx applied to both destroys it. Radians are the other trap; sin 5° ≈ 5 works only because 5° = 0.087 rad, and students who plug the degree number itself get answers off by a factor of 57. Finally, approximations are for eliminating options and estimating, not for the final digit — when two options differ by 2 percent, carry one more term than feels necessary.
Frequently asked questions
When is the approximation (1 + x)^n ≈ 1 + nx valid?
Only when |x| is much smaller than 1 and n is finite; if x approaches 1 or the required precision is high, retain the n(n−1)x²/2 term as well.
In what units must θ be for sin θ ≈ θ?
Radians exclusively; sin 5° ≈ 0.087 works because 5 degrees is 0.087 radians, whereas inserting 5 gives an error of a factor near 57.
How do percentage changes propagate through a formula?
Take logarithms and differentiate: each exponent multiplies the relative change of its variable, so T ∝ √l means ΔT/T = (1/2)(Δl/l).
Why keep the θ²/2 term in cos θ?
Because in cancellation problems the leading 1 drops out, and the θ²/2 difference is the entire physical effect — first-order truncation leaves zero.
How can limiting cases check an answer?
Push a mass to infinity, a distance to zero, or an angle to 90 degrees and confirm the formula reproduces a known result; any mismatch flags an algebra error before you commit an option.