In short

The honest answer is both — and not in a vague "balance" way. Memorise the 5-7 most-used identities so your hand writes a^2 - b^2 = (a+b)(a-b) in the same time it takes to read it. That is raw exam speed and it is non-negotiable for JEE-paced papers. But also learn to derive every one of them from FOIL or the distributive law. Derivation is your safety net: if you ever blank on (a-b)^3 in the middle of a problem, you can rebuild it in twenty seconds. Without the derivation skill, a forgotten identity is a lost question. With it, you are never stuck. Why: memory gives you speed on familiar shapes; derivation gives you survival on unfamiliar ones. You need both because exams give you both.

You have probably had this argument with yourself at 11pm. You are revising algebra. You see (a+b)^3 = a^3 + 3a^2b + 3ab^2 + b^3 in your notes. The honest question hits: should you memorise this thing, or should you just rederive it every time from (a+b)(a+b)(a+b)?

Your tuition teacher said memorise. Your school sir said understand. A YouTube channel told you "real mathematicians don't memorise". Your seniors said "in JEE you have no time, just memorise". Everyone is partly right. This article gives you the whole picture, with a balance scale you can keep in your head.

A balance scale with two pans labelled memory and derivation skillA balance scale with a central pivot. The left pan is labelled "MEMORY — speed" and holds a stack of small cards labelled with identities. The right pan is labelled "DERIVATION — safety net" and holds a small toolbox labelled FOIL. Both pans are at the same level, indicating both are essential for stable algebra performance. a²−b² FOIL MEMORY speed in exams DERIVATION safety net stable algebra performance
Both pans must be loaded. An empty memory pan slows you down on every familiar identity; an empty derivation pan leaves you stranded the moment something unfamiliar — or something you blanked on — appears. Strong students keep both stocked.

The case for memorising

Memorising the standard identities is not lazy. It is what gives you the speed that hard exams demand.

Imagine the JEE Mains paper. You have on average about a minute per question. You see x^4 - 16 inside a larger problem and you need it factored now. If you have memorised a^2 - b^2 = (a+b)(a-b), your eye sees this:

x^4 - 16 = (x^2)^2 - 4^2 = (x^2 - 4)(x^2 + 4) = (x-2)(x+2)(x^2+4)

Two passes of difference of squares, six seconds, done. Why: pattern recognition is instantaneous when the pattern is in long-term memory. You do not "compute" the answer — you recognise it, the way you recognise a friend's face.

The student who refuses to memorise has to expand (x^2-4)(x^2+4) on rough paper to verify it equals x^4 - 16. They get there, but they spent thirty seconds. Across a 90-question paper that is forty-five extra minutes — which they do not have.

There are three concrete benefits to memorisation:

  1. Speed on familiar shapes. Identities recognised by reflex are essentially free. The arithmetic you would have done is gone.
  2. Pattern recognition. Once a^2 - b^2 is burned in, you start seeing it everywhere — inside \sin^2 x - \cos^2 x, inside 1 - p^2, inside x^4 - 1 (twice!). You see structure that an un-memorised student misses.
  3. Fewer arithmetic slips. A memorised identity has no expansion step where you can drop a sign or forget a cross term. The cross terms are baked into the formula.

Here is the essentials list every Indian school student should have in long-term memory by class 10:

That is eight identities. Most JEE aspirants get them into memory using spaced repetition — flashcard apps like Anki, or simple physical flashcards reviewed at 1 day, 3 days, 7 days, 21 days. The spacing is the trick: a card you almost forgot, then re-saw, sticks for months. This is the same technique medical students use for anatomy, and it works just as well for algebra.

The student who memorised but never derived

Riya memorised (a+b)^2 = a^2 + 2ab + b^2 in class 8. She got full marks on every quiz that asked her to expand it. In class 9, on a problem where she had to expand (x + 5)^2 in a hurry, her hand wrote — confidently — (x+5)^2 = x^2 + 25.

She had over-trusted memory and never built the FOIL habit. Without the derivation reflex, her brain pattern-matched the squared bracket onto the wrong template (the distributive k(a+b) = ka + kb shape) and she could not feel that it was wrong. She lost two marks.

Why: pure memory without understanding is brittle. Under fatigue or speed, the wrong card surfaces and you have no way to check it. Derivation is the check.

The case for deriving

Now the other pan. Why not just memorise everything and skip derivation?

Because exams are not made of only the eight identities you memorised. Eventually you meet:

In every one of these, memory alone fails. Derivation is what gets you through.

Concretely, the derivation skill gives you four things:

  1. Recovery from forgetting. Blank on a^3 - b^3? Multiply (a-b)(a^2 + ab + b^2) in twenty seconds and confirm it equals a^3 - b^3. Identity restored without panic.
  2. Spotting non-standard variants. Sophie Germain's identity says a^4 + 4b^4 = (a^2 + 2b^2 - 2ab)(a^2 + 2b^2 + 2ab). No school memorises this. But a student who knows how to factor by clever grouping can derive it on the spot when a problem demands it.
  3. Algebra muscle. Every derivation is a rep at FOIL, sign tracking, collecting like terms. These are the underlying skills that show up in every algebra problem, not just identity questions. The strong derivation student is also strong at quadratics, partial fractions, and polynomial division.
  4. Generalising fearlessly. Once you have derived (a+b)^2, (a+b)^3, (a+b)^4 you start seeing Pascal's triangle and the binomial theorem. Memorisers stop at (a+b)^3; derivers see the pattern continue forever.

There is a deeper reason, too. In the long run, the algebraic engine outweighs flashcard memory. A class 10 student survives on memorised identities. A JEE aspirant needs to handle expressions their school never showed them. A college student in an engineering or physics course meets identities for matrices, vectors, complex numbers — none of which are on any memorised list. The students who built the derivation engine just keep going. The ones who relied only on memory hit a wall.

The student who can always derive

Arjun never bothered to memorise (a+b)^4. In a JEE-level problem he meets (2x + 5y)^4. He calmly writes Pascal's triangle row 4 in the margin: 1, 4, 6, 4, 1. Then:

(2x+5y)^4 = (2x)^4 + 4(2x)^3(5y) + 6(2x)^2(5y)^2 + 4(2x)(5y)^3 + (5y)^4
= 16x^4 + 160 x^3 y + 600 x^2 y^2 + 1000 x y^3 + 625 y^4

He took forty-five seconds — slower than a student with a memorised binomial expansion — but he got it correct, no formula sheet needed, and he could equally have done (3a-2b)^5 or any other binomial. The engine scales.

Why: derivation skill is general. It does not care whether the exponent is 2, 3, 4 or 7, or whether the inside is a+b or 2x+5y. Memory is specific; derivation is universal.

The right balance

You do not have to choose. The strongest students do this:

  1. Memorise the form of the eight standard identities. Aim for instant recognition — see a^2 - b^2, write (a+b)(a-b) without thinking.
  2. Derive each identity 3–4 times by hand during your first study of the chapter. Use FOIL or the geometric square picture. Watch every cross term appear. Feel why 2ab is there and not ab.
  3. Re-derive once a month as a sanity check. Spaced repetition for the form, occasional re-derivation for the understanding. Both pans stay loaded.
  4. For any non-standard expression, fall back to derivation. Never force a memorised identity onto a shape it does not fit (that is exactly the misconception trap from (a+b)^2).

In short: memory for speed, derivation for safety. Each compensates for the other's weakness.

The ideal — multiple paths

Sneha sees x^6 - 1 in a problem. Her trained eye notices two patterns at once.

Path A — difference of squares: x^6 - 1 = (x^3)^2 - 1^2 = (x^3 - 1)(x^3 + 1). Then expand each cube factor:

= (x-1)(x^2+x+1)(x+1)(x^2-x+1)

Path B — difference of cubes: x^6 - 1 = (x^2)^3 - 1^3 = (x^2 - 1)((x^2)^2 + x^2 + 1). Then:

= (x-1)(x+1)(x^4 + x^2 + 1)

Both are correct. Path A is fully factored over the rationals; path B leaves a quartic that further factors into the same pieces. She picks whichever is more useful for the next step. Why: only the student who has both the memorised difference-of-squares pattern AND the derivation flexibility can see both paths. Pure memorisers see only one. Pure derivers see neither without doing the work.

That is the goal. Eight identities at your fingertips, FOIL as your fallback engine, and the judgement to know when each is the right tool.

References

  1. NCERT Class 9 Mathematics, Chapter 2: Polynomials — the standard derivations of the school identities.
  2. Wikipedia: Sophie Germain identity — a non-standard identity you can only handle by deriving.
  3. Wikipedia: Pascal's triangle — the engine for any binomial expansion.
  4. Anki — open-source spaced repetition — the tool most JEE aspirants use to keep facts in long-term memory.
  5. Wikipedia: Spaced repetition — the cognitive-science backing for why flashcard review at growing intervals beats cramming.
  6. Khan Academy: Polynomial identities — practice that mixes memorised use with derivation.