Algebraic identities: squares and difference of squares

Multiply 53×4753 \times 47 in your head. Long multiplication takes half a minute and invites arithmetic slips — but notice that 5353 and 4747 sit three steps above and below the round number 5050. That symmetry is not a coincidence you have to spot fresh every time; it is a pattern with a name, an algebraic identity, and once you know it the answer drops out in seconds as 502−32=249150^2 - 3^2 = 2491. This article covers the three standard identities of Class 8: the square of a sum (a+b)2(a+b)^2, the square of a difference (a−b)2(a-b)^2, and the difference of squares (a+b)(a−b)(a+b)(a-b). Each is not a rule to memorise blindly but a direct consequence of the distributive law you already know from expanding two binomials. Once derived and understood, they turn awkward-looking squares and products — 9972997^2, 998×1002998 \times 1002, 8.928.9^2 — into a few lines of easy arithmetic, and they set up factorisation, a core Class 8–9 skill, in the chapter ahead.

The core insight

The three products (a+b)2(a+b)^2, (a−b)2(a-b)^2 and (a+b)(a−b)(a+b)(a-b) expand, by the distributive law, to a2+2ab+b2a^2+2ab+b^2, a2−2ab+b2a^2-2ab+b^2 and a2−b2a^2-b^2 respectively, and recognising which pattern a given multiplication matches turns slow arithmetic into a few quick steps.

The core idea

An identity is an equation true for every value of the letters in it — unlike an ordinary equation, which is only true for particular values and has to be solved. The three identities below are always true, for any aa and bb whatsoever, which is exactly why they are worth memorising and applying on sight.

1. Square of a sum. Expanding (a+b)(a+b)(a+b)(a+b) by the distributive law:

(a+b)2=a2+2ab+b2(a+b)^2 = a^2 + 2ab + b^2

where: aa and bb are the two terms of the bracket, a2a^2 and b2b^2 are their individual squares, and 2ab2ab is twice their product — the "middle term" that is easiest to forget.

2. Square of a difference. Expanding (a−b)(a−b)(a-b)(a-b) the same way, but with a subtraction:

(a−b)2=a2−2ab+b2(a-b)^2 = a^2 - 2ab + b^2

where: the last term is still +b2+b^2, never −b2-b^2, because a negative times a negative is positive; only the middle term changes sign compared to the identity above.

3. Difference of squares. Expanding (a+b)(a−b)(a+b)(a-b), the two middle terms are opposites and cancel out completely rather than combining:

(a+b)(a−b)=a2−b2(a+b)(a-b) = a^2 - b^2

where: the result has only two terms, not three — this is the tell-tale sign that a product matches this pattern rather than one of the two square identities above. The figure below shows why (a+b)2(a+b)^2 has that extra 2ab2ab term: it is a geometric picture, not just algebra. A square of side (a+b)(a+b) splits into a square of side aa, a square of side bb, and two identical rectangles of area abab each — hence 2ab2ab.

Figure
A large square of side a plus b divided into a smaller a by a square a b by b square and two equal a by b rectangles whose combined area is two a b
Why (a+b)² = a² + 2ab + b² a² ab ab b² a b a b Total area = a² + ab + ab + b² = a² + 2ab + b² The two ab rectangles are why the middle term is doubled.

The two square identities differ only in the sign of the middle term 2ab2ab; the difference-of-squares identity is structurally different, collapsing to two terms because its middle terms cancel. Spotting which of the three shapes a product matches — same bracket twice with a plus, same bracket twice with a minus, or a plus bracket paired with a minus bracket of the same two terms — is the key skill before applying any identity.

Derivation ADeriving all three identities from the distributive law

The goal is to show that all three standard identities are not separate facts to memorise but a single idea, the distributive law x(y+z)=xy+xzx(y+z) = xy + xz, applied twice to a product of two binomials — so that if you ever forget an identity in an exam, you can rebuild it in under a minute from first principles.

Step 1. Start from the general binomial product, which you already know how to expand: (a+b)(c+d)=ac+ad+bc+bd(a+b)(c+d) = ac + ad + bc + bd, obtained by distributing each term of the first bracket over the second.

Step 2. For the square of a sum, set c=ac=a and d=bd=b, so the product becomes (a+b)(a+b)(a+b)(a+b). Substituting into the general expansion: (a+b)(a+b)=a⋅a+a⋅b+b⋅a+b⋅b=a2+ab+ab+b2(a+b)(a+b) = a\cdot a + a\cdot b + b\cdot a + b\cdot b = a^2 + ab + ab + b^2. Since ab=baab = ba, the two middle terms are identical and combine: a2+2ab+b2a^2 + 2ab + b^2.

Step 3. For the square of a difference, set c=ac=a and d=−bd=-b, so the product is (a−b)(a−b)(a-b)(a-b). The same distribution gives a2−ab−ab+b2a^2 - ab - ab + b^2, since (−b)×(−b)=+b2(-b)\times(-b) = +b^2. The two middle terms are again identical (both −ab-ab) and combine to give a2−2ab+b2a^2 - 2ab + b^2.

Step 4. For the difference of squares, keep the first bracket as (a+b)(a+b) but pair it with (a−b)(a-b) instead, i.e. c=ac=a, d=−bd=-b in one bracket only: (a+b)(a−b)=a⋅a−a⋅b+b⋅a−b⋅b=a2−ab+ab−b2(a+b)(a-b) = a\cdot a - a\cdot b + b\cdot a - b\cdot b = a^2 - ab + ab - b^2. This time the middle terms are opposites, −ab-ab and +ab+ab, and they cancel to zero, leaving only a2−b2a^2 - b^2. All three results depend on nothing beyond one application of the distributive law, so they hold for any real numbers aa and bb — which is exactly what makes them identities rather than equations to be solved.

Worked example

Use the difference-of-squares identity to compute 998×1002998 \times 1002 without long multiplication, and check the result against the earlier example of 9972997^2.

Step 1. Notice that 998998 and 10021002 are equally spaced around the round number 10001000: 998=1000−2998 = 1000 - 2 and 1002=1000+21002 = 1000 + 2.

Step 2. Match this to the pattern (a+b)(a−b)(a+b)(a-b) with a=1000a=1000, b=2b=2, so apply the identity (a+b)(a−b)=a2−b2(a+b)(a-b) = a^2 - b^2:

998×1002=(1000−2)(1000+2)=10002−22998 \times 1002 = (1000-2)(1000+2) = 1000^2 - 2^2

Step 3. Compute each piece:

10002=1,000,000,22=41000^2 = 1{,}000{,}000, \qquad 2^2 = 4

Step 4. Subtract:

998×1002=1,000,000−4=999,996998 \times 1002 = 1{,}000{,}000 - 4 = 999{,}996

The product is 998×1002=999,996998 \times 1002 = 999{,}996. As a check, this should be close to but slightly less than 10002=1,000,0001000^2 = 1{,}000{,}000, and 999,996999{,}996 fits: the "loss" from 1,000,0001{,}000{,}000 is exactly b2=4b^2 = 4, however far apart aa and bb are chosen, showing the same shortcut used earlier for 9972997^2 and 53×4753\times 47 works identically here.

Competitive stretch

The same trick extends to decimals: 8.928.9^2 is (9−0.1)2=92−2(9)(0.1)+0.12=81−1.8+0.01=79.21(9-0.1)^2 = 9^2 - 2(9)(0.1) + 0.1^2 = 81 - 1.8 + 0.01 = 79.21 — no long multiplication of decimals needed, only the square-of-a-difference identity with a=9a=9, b=0.1b=0.1.

Common traps

  • Forgetting the middle term. Writing (a+b)2=a2+b2(a+b)^2 = a^2+b^2 drops the 2ab2ab entirely; squaring a sum is not the same as summing the squares.
  • Sign confusion in (a−b)2(a-b)^2. The last term stays +b2+b^2, never −b2-b^2; only the middle term flips sign compared to (a+b)2(a+b)^2.
  • Forcing the wrong identity. The difference-of-squares pattern needs one bracket with a plus and one with a minus of the same two terms; it cannot be applied to (a+b)(a+b)(a+b)(a+b) or (a−b)(a−b)(a-b)(a-b).
  • Choosing an awkward split. The shortcut only saves time if aa is round and easy to square, like 5050, 100100 or 10001000; splitting 998998 as (500+498)(500+498) instead of (1000−2)(1000-2) throws away the advantage.

Ask the tutor: pick any three-digit number close to a multiple of 100 and ask the tutor to walk through the fastest identity-based way to square it.

Why do I need to memorise these identities if I can just multiply the brackets out every time?

You can always fall back on the distributive law, and in fact that is exactly how you should rebuild an identity if you forget it — Step 1 to Step 4 of the derivation above takes under a minute. But in a timed exam, recognising the pattern instantly and writing the three-term or two-term answer directly is much faster than expanding (a+b)(a+b)(a+b)(a+b) term by term every single time, especially for numeric shortcuts like 9972997^2 or 998×1002998\times1002 where speed matters.

How do I know whether to use the square identity or the difference-of-squares identity on a given multiplication?

Look at the two brackets. If both brackets are literally the same, like (a+b)(a+b)(a+b)(a+b) or (a−b)(a−b)(a-b)(a-b), you have a square — use (a+b)2(a+b)^2 or (a−b)2(a-b)^2 and expect three terms. If the two brackets share the same two terms but one has a plus and the other a minus, like (a+b)(a−b)(a+b)(a-b), you have a difference of squares — expect only two terms, since the middle terms cancel. For a numeric shortcut, check whether the two numbers are equally spaced above and below a round middle value; if so, that middle value is aa and the spacing is bb.

Does the order of a and b matter, and can b be negative?

The identities hold for any real numbers aa and bb, including negative ones, because the derivation only used the distributive law, which works for all real numbers. Swapping aa and bb does not change (a+b)2(a+b)^2 or (a+b)(a−b)(a+b)(a-b) in the sense that the set of values used is the same, but it does change which number plays which role — always match aa to the round, easy-to-square number when using the identities as a mental-math shortcut, since that is what makes the calculation fast.

Before this

Where this leads