Symmetric Expressions: When You See x+y and xy, Think Identities — Not Brute Force

In short

When a problem gives you x + y and xy (or x - y and xy, or x - 1/x, or any pair of "symmetric inputs"), and asks you to compute x^2 + y^2, x^3 + y^3, x^3 - y^3, or similar — your toolkit is the symmetric identities, not the quadratic formula. The bridges x^2 + y^2 = (x+y)^2 - 2xy, x^3 + y^3 = (x+y)^3 - 3xy(x+y), and (x+y)^2 - (x-y)^2 = 4xy let you compute the answer in one line. Brute-force solving for x and y separately almost always takes ten times longer.

You open a JEE problem. It says: "If x + y = 5 and xy = 6, find x^2 + y^2." Your first instinct might be to solve for x and y — set up the quadratic t^2 - 5t + 6 = 0, factor it, get t = 2 or t = 3, then compute 2^2 + 3^2 = 4 + 9 = 13.

That works. It also takes forty seconds.

The smarter move takes five seconds. Notice that (x+y)^2 = x^2 + 2xy + y^2. Rearrange: x^2 + y^2 = (x+y)^2 - 2xy = 25 - 12 = 13. Done. You never needed to know what x and y individually were — you only needed how they combine.

That is the heuristic this article is about. When the problem speaks in symmetric quantities — x+y, xy, x-y, x \pm 1/x — the answer wants symmetric identities, not individual values. Train your eye to spot the pattern, and most "find x^n \pm y^n" problems collapse from a calculation into a substitution.

The toolkit — four bridges to memorise

These four identities are the entire game. Every "given x+y and xy, find …" problem is one of these in disguise.

The symmetric toolkit. Once you can read these four identities at a glance, almost every "given $x+y$ and $xy$" problem reduces to a substitution. The blue $+$ and red $-$ are there to help you spot which version of the identity you need.

A few notes on each:

x^2 + y^2 = (x+y)^2 - 2xy. Expand (x+y)^2 = x^2 + 2xy + y^2 and subtract 2xy. The 2xy is the "cross-term tax" you pay when squaring a sum. Why: any square of a sum produces an interaction term. Subtracting 2xy pulls it back out, leaving the pure sum of squares.
x^3 + y^3 = (x+y)^3 - 3xy(x+y). Expand (x+y)^3 = x^3 + 3x^2y + 3xy^2 + y^3 = x^3 + y^3 + 3xy(x+y). Subtract 3xy(x+y) to isolate x^3 + y^3. You can also factor it as (x+y)\bigl((x+y)^2 - 3xy\bigr), which is sometimes more convenient.
x^3 - y^3 = (x-y)^3 + 3xy(x-y). Same logic with a sign flip — (x-y)^3 = x^3 - 3x^2y + 3xy^2 - y^3 = x^3 - y^3 - 3xy(x-y). Solve for x^3 - y^3 and you add the 3xy(x-y) correction.
(x+y)^2 - (x-y)^2 = 4xy. Expand both squares; the x^2 and y^2 terms cancel, the cross-terms add up. This is the "polarisation identity" of school algebra — it lets you recover xy from sum and difference.

Why these four exist together: they are reformulations of the same underlying fact — every symmetric polynomial in x and y can be written using only x+y and xy. Those two are called the elementary symmetric polynomials, and the toolkit above is just the first few cases of a general theorem (more on that at the end).

Worked example 1 — sum of squares

Given x + y = 5 and xy = 6, find x^2 + y^2.

Reach for x^2 + y^2 = (x+y)^2 - 2xy:

x^2 + y^2 = 5^2 - 2 \cdot 6 = 25 - 12 = 13

Done. One line. The brute-force approach (solve t^2 - 5t + 6 = 0 to get t = 2, 3, then compute 4 + 9 = 13) gives the same answer but takes ten times longer — and would not work at all if the discriminant were ugly.

Sum and product feed in; sum of squares falls out. No detour through individual values.

Worked example 2 — sum of cubes

Given x + y = 3 and xy = 1, find x^3 + y^3.

The identity is x^3 + y^3 = (x+y)^3 - 3xy(x+y). Substitute:

x^3 + y^3 = 3^3 - 3 \cdot 1 \cdot 3 = 27 - 9 = 18

Try the brute-force route and you'll see why this matters. The quadratic t^2 - 3t + 1 = 0 has discriminant 9 - 4 = 5, giving t = \frac{3 \pm \sqrt{5}}{2} — irrational roots. To compute x^3 + y^3 from those you'd need to cube two ugly surds and add them. The symmetric route doesn't care that the roots are irrational; it operates entirely on the sum and the product.

The roots are irrational. Doesn't matter — the symmetric route ignores them.

Worked example 3 — the x \pm 1/x family

Given x - \dfrac{1}{x} = 4, find x^3 - \dfrac{1}{x^3}.

This is the same trick wearing a costume. Set y = 1/x — then xy = x \cdot (1/x) = 1, and you're given x - y = 4. You want x^3 - y^3.

Use x^3 - y^3 = (x-y)^3 + 3xy(x-y):

x^3 - \tfrac{1}{x^3} = 4^3 + 3 \cdot 1 \cdot 4 = 64 + 12 = 76

Or, equivalently, cube both sides of x - 1/x = 4 and rearrange. (x - 1/x)^3 = x^3 - 3x^2 \cdot (1/x) + 3x \cdot (1/x^2) - 1/x^3 = x^3 - 1/x^3 - 3(x - 1/x). So:

64 = x^3 - \tfrac{1}{x^3} - 3 \cdot 4 \quad\Rightarrow\quad x^3 - \tfrac{1}{x^3} = 64 + 12 = 76

Same answer. The "x \pm 1/x" pattern is everywhere in JEE algebra precisely because the product x \cdot (1/x) collapses to 1, leaving you with a clean symmetric problem in disguise. Why: hiding xy = 1 inside the substitution lets the question-setter ask one quantity (x - 1/x) and demand another (x^3 - 1/x^3) without ever giving you x. The symmetric identity is the only way through.

How to spot the pattern

The cue is almost always one of these:

The problem gives you x+y and xy (or x-y and xy).
The problem asks for x^2 + y^2, x^2 - y^2, x^3 \pm y^3, x^4 + y^4, etc.
The problem features x \pm 1/x — which is just the symmetric problem with y = 1/x secretly imposed.
The problem mentions the roots of a quadratic t^2 + pt + q = 0. By Vieta's relations, the roots \alpha, \beta satisfy \alpha + \beta = -p and \alpha\beta = q — exactly the symmetric inputs.

If any of these light up, your first move should be reaching for the toolkit, not the quadratic formula. The roots themselves are usually a distraction. The symmetric quantities are what the problem is really about.

Higher powers — climb the staircase

What about x^4 + y^4 or x^5 + y^5? You don't need new identities — you bootstrap from the lower ones. For instance:

x^4 + y^4 = (x^2 + y^2)^2 - 2(xy)^2

Compute x^2 + y^2 first using the toolkit, then square it and subtract 2(xy)^2. Or use Newton's identity: x^{n+1} + y^{n+1} = (x+y)(x^n + y^n) - xy(x^{n-1} + y^{n-1}). This recurrence climbs powers one at a time, using only x+y and xy. So given the pair (x+y, xy), every power sum x^n + y^n is computable — never once needing the individual values.

The deeper reason — elementary symmetric polynomials

This isn't a bag of tricks. It is the first appearance, in school maths, of one of the most beautiful theorems in algebra: the fundamental theorem of symmetric polynomials. It says that every polynomial expression in x and y that is unchanged when you swap x and y — that is, every symmetric polynomial — can be written using only the two elementary symmetric polynomials e_1 = x + y and e_2 = xy.

So x^2 + y^2, x^3 + y^3, x^4 + y^4 — none of them require knowing x and y separately. They are functions of e_1 and e_2 alone. The four bridges in the toolkit are the small change you make to translate.

For three variables, the elementary symmetric polynomials are e_1 = x+y+z, e_2 = xy+yz+zx, e_3 = xyz — exactly the coefficients you see in Vieta's formulas for cubic equations. The same machinery powers everything from school algebra to Galois theory. The fact that x^2 + y^2 = (x+y)^2 - 2xy is, in disguise, a profound piece of mathematics — you are watching symmetric polynomials being expressed in their natural basis.

The takeaway

The instant you see x + y and xy in a problem statement — or x \pm 1/x, or roots of a quadratic — your hand should reach for the symmetric identities, not the quadratic formula. Sum and product are the two coordinates that fully describe what the problem cares about. Solving for x and y individually is doing extra work the problem never required.

Memorise the four bridges. Recognise the pattern. Substitute. Done in one line.