You're stuck. The polynomial in front of you has four terms, but grouping doesn't work. Or three terms, but split-the-middle refuses to land on integers. Or it's degree 4 or 5, and none of the standard identities — difference of squares, perfect square, sum or difference of cubes — fit.

Before you give up, there is one more tool. The Rational Root Theorem does not factor the polynomial for you. It gives you a finite list of candidate roots to test. If the polynomial has any rational root, it must be on that list. Test each; if one gives P(x) = 0, you have found a factor.

The theorem

Let P(x) = a_n x^n + a_(n-1) x^(n-1) + ... + a_1 x + a_0 be a polynomial with integer coefficients. If P(x) has a rational root p/q written in lowest terms (so gcd(p, q) = 1), then:

So the full set of candidate roots is

\text{candidates} = \pm \frac{\text{factors of } a_0}{\text{factors of } a_n}.

This is a finite list. You write it out, and you test each one.

Why it works

Suppose p/q is a root, with gcd(p, q) = 1. Then

a_n \left(\tfrac{p}{q}\right)^n + a_{n-1} \left(\tfrac{p}{q}\right)^{n-1} + \cdots + a_0 = 0.

Multiply through by q^n to clear denominators:

a_n p^n + a_{n-1} p^{n-1} q + a_{n-2} p^{n-2} q^2 + \cdots + a_0 q^n = 0.

Move a_0 q^n to the other side:

a_n p^n + a_{n-1} p^{n-1} q + \cdots + a_1 p q^{n-1} = -a_0 q^n.

Every term on the left has a factor of p, so p \mid a_0 q^n. Since gcd(p, q) = 1, p must divide a_0.

Now move a_n p^n to the other side instead:

a_{n-1} p^{n-1} q + \cdots + a_0 q^n = -a_n p^n.

Every term on the left has a factor of q, so q \mid a_n p^n. Since gcd(p, q) = 1, q divides a_n. Done.

Worked example 1 — monic cubic

Take P(x) = x³ - 6x² + 11x - 6.

Leading coefficient is 1. Constant term is -6. So q must divide 1, meaning q = 1. And p must divide 6, meaning p ∈ {1, 2, 3, 6}. Including signs, the candidates are

\pm 1, \pm 2, \pm 3, \pm 6.

Test the smallest first. P(1) = 1 - 6 + 11 - 6 = 0. Found one. So (x - 1) is a factor.

Divide: x³ - 6x² + 11x - 6 \div (x - 1) = x² - 5x + 6. The quotient is a quadratic, which factors as (x - 2)(x - 3).

Full factorisation:

P(x) = (x - 1)(x - 2)(x - 3).

Roots: 1, 2, 3. All three were on the candidate list.

Worked example 2 — non-monic cubic

Now P(x) = 2x³ + 3x² - 11x - 6.

Leading coefficient is 2. Constant is -6. So q ∈ {1, 2} and p ∈ {1, 2, 3, 6}. The candidate set is

\pm \left\{1,\, 2,\, 3,\, 6,\, \tfrac{1}{2},\, \tfrac{3}{2}\right\}.

Twelve candidates. Test small integers first. P(1) = 2 + 3 - 11 - 6 = -12. P(-1) = -2 + 3 + 11 - 6 = 6. P(2) = 16 + 12 - 22 - 6 = 0. Got it.

So (x - 2) is a factor. Dividing,

2x³ + 3x² - 11x - 6 = (x - 2)(2x² + 7x + 3).

The quadratic 2x² + 7x + 3 factors as (2x + 1)(x + 3). So

P(x) = (x - 2)(2x + 1)(x + 3).

Roots: 2, -1/2, -3. Notice the root -1/2 was on the candidate list as -1/2.

What if no rational root exists

You test every candidate; none gives P = 0. That means the polynomial has no rational roots — not that it has no roots. It may still have irrational real roots (like \sqrt{2} or \sqrt[3]{5}) or complex roots.

For irrational real roots you fall back on numerical methods (bisection, Newton's method). For most school and JEE problems, though, the polynomial is designed to have a rational root. The Rational Root Theorem is built for exactly this case.

Pragmatic strategy

Worked example 3 — Rational Root Theorem finds nothing useful

P(x) = x² - 2x + 2.

Discriminant: (-2)² - 4(1)(2) = 4 - 8 = -4. Negative — no real roots at all. But suppose you didn't notice that and tried the Rational Root Theorem anyway. Candidates: ±1, ±2.

Nothing is zero. The theorem correctly tells you there are no rational roots. The quadratic formula then gives complex roots 1 \pm i.

The lesson: the theorem can confirm a negative result. If no candidate works, no rational root exists.

Use with synthetic division

Once you find a candidate a with P(a) = 0, synthetic division extracts the quotient cleanly. Write the coefficients, bring down the first, multiply by a, add to the next, repeat. The result is the quotient polynomial, one degree lower.

Then recurse. Apply the Rational Root Theorem to the quotient. Find another root. Divide again. Continue until you reach a quadratic, which you finish by formula or by sight.

Limits of the theorem

Three things to remember.

Related tool — Descartes' Rule of Signs

A useful companion. Count the number of sign changes in the sequence of coefficients of P(x). The number of positive real roots is at most that count, and differs from it by an even number. Do the same for P(-x) to bound the number of negative real roots.

For P(x) = x³ - 6x² + 11x - 6, the signs are +, -, +, - — three sign changes, so up to three positive real roots. P(-x) = -x³ - 6x² - 11x - 6 has signs -, -, -, - — zero sign changes, so no negative real roots. You can skip testing the negative candidates entirely.

Recognition drill

For each polynomial, list all rational root candidates.

The shorter the list, the faster the search.

Common confusions

Closing

When factoring stalls and identities don't fit, the Rational Root Theorem hands you a finite menu. Test each candidate; most school and exam problems yield within the first three or four tries. Once you have one root, synthetic division knocks the degree down, and you recurse. Irrational roots slip through — but for the integer-coefficient polynomials in your textbook, it is the reliable last resort.