You sit down to factor 2x² + 7x + 3. You remember the trick: "find two numbers that multiply to give the constant and add to give the middle." Factor pairs of 3: (1, 3), (-1, -3). Sums: 4 and -4. Neither is 7. You write "doesn't factor" and move on.

But it does factor. Cleanly. Into (2x + 1)(x + 3). The trick you remembered was a special case of the right trick that only works when the leading coefficient is 1. The moment the leading coefficient is anything else, you need a different starting number. Not c. The product ac.

The "split the middle" procedure, correctly stated

For ax² + bx + c:

  1. Compute ac — the product of the leading coefficient and the constant. Not c alone.
  2. Find two integers p and q with p + q = b and p × q = ac.
  3. Split the middle term: rewrite as ax² + px + qx + c.
  4. Group the four terms in pairs. Pull out the common factor from each pair.
  5. If step 1 was right, both pairs leave behind the same bracket. Pull that bracket out.

Every line depends on step 1 being right. Use c instead of ac and you will not find numbers that work — or you will find numbers that look like they work but produce nonsense at the grouping stage.

Worked example — 2x² + 7x + 3

a = 2, b = 7, c = 3. So ac = 6.

You need two integers with product 6 and sum 7. Factor pairs of 6: (1, 6), (2, 3). The pair (1, 6) sums to 7.

Split: 2x² + 6x + x + 3.

Group: (2x² + 6x) + (x + 3) = 2x(x + 3) + 1(x + 3).

The bracket (x + 3) is common. Pull it out:

(2x + 1)(x + 3)

Verify: (2x + 1)(x + 3) = 2x² + 6x + x + 3 = 2x² + 7x + 3. Correct.

Using c alone (wrong) — why it fails

Retrace with c = 3 instead of ac = 6. You need integers with product 3 and sum 7. Factor pairs of 3: (1, 3) and (-1, -3). Sums: 4 and -4. Neither is 7.

A student following the wrong rule concludes 2x² + 7x + 3 does not factor over the integers. That is false. The polynomial factors perfectly. The method failed, not the polynomial.

This is the dangerous part of the misconception. It does not announce itself with an absurd answer. It quietly tells you "no integer pair exists" when one does. You walk away thinking the problem was harder than it was.

Why ac, not c — the proof

Write ax² + bx + c = (αx + β)(γx + δ). Expand the right side:

(αx + β)(γx + δ) = αγx² + (αδ + βγ)x + βδ

Matching coefficients with ax² + bx + c:

"Split the middle" means writing b as αδ + βγ — two pieces. So you want two numbers p = αδ and q = βγ with:

The product of the split numbers is ac, not c. The number c alone is βδ, only one of the four products in the expansion. ac is what falls out naturally.

When a = 1, ac = c, so the two rules coincide. That is the only reason the "just c" trick ever worked. It was a consequence of a = 1, not a general rule.

Another worked example — 6x² - 13x + 5

a = 6, b = -13, c = 5. ac = 30.

You need integers with product 30 and sum -13. Sum negative, product positive, so both numbers are negative. Negative factor pairs of 30: (-1, -30), (-2, -15), (-3, -10), (-5, -6). Sums: -31, -17, -13 — the pair is -3 and -10.

Split: 6x² - 10x - 3x + 5.

Group: (6x² - 10x) + (-3x + 5) = 2x(3x - 5) - 1(3x - 5).

(3x - 5) is common:

(2x - 1)(3x - 5)

Verify: (2x - 1)(3x - 5) = 6x² - 10x - 3x + 5 = 6x² - 13x + 5. Correct.

When "just c" works — only if a = 1

For monic x² + bx + c, the leading coefficient is 1, so ac = c. Both rules give the same numbers.

Take x² + 7x + 12. ac = 12. Pair (3, 4) works: (x + 3)(x + 4). The "just c" answer is the same.

This is the only case where "just c" is safe. Outside it, you are gambling.

The full decision

No penalty for using ac when a = 1. Heavy penalty for using c when a ≠ 1. The safe habit is universal: write ac = at the top of your scratch, every time.

When the polynomial doesn't factor

Sometimes no integer pair with product ac and sum b exists. Then the polynomial does not factor over the rationals — it has irrational or complex roots. Switch to the quadratic formula.

Example: 2x² + 3x + 4 has ac = 8, b = 3. Pairs of 8: (1, 8), (2, 4). Sums: 9, 6. Negative pairs give negative sums. No integer split exists. The discriminant b² - 4ac = 9 - 32 = -23 is negative, confirming complex roots.

The point: only conclude "no integer factorisation" after searching factor pairs of ac, not c. Concluding too early — based on c alone — is the misconception this page is trying to kill.

Common error pattern

Try 6x² + 5x - 4. a = 6, b = 5, c = -4. ac = -24.

The wrong rule looks at factor pairs of -4: (-1, 4), (-2, 2), (-4, 1). Sums never reach 5. Student writes "doesn't factor."

Correct rule: ac = -24. Try (-3, 8): product -24, sum 5.

Split: 6x² - 3x + 8x - 4 = 3x(2x - 1) + 4(2x - 1) = (3x + 4)(2x - 1). The polynomial factors. The student who used c alone never had a chance.

Alternative name — "ac decomposition"

Some textbooks call this the "ac method": "multiply a and c, find factors of that product summing to b, rewrite the middle, group." Same procedure, more honest name — it reminds you ac is what to compute. If your textbook says "split the middle" without mentioning ac, mentally translate to "ac decomposition."

Recognition drill

Run through these. For each, compute ac, find the pair, split, group, factor.

In every case the first move is the same: compute ac. The rest then has somewhere to start.

Closing

The method needs two numbers with product ac and sum b. Not c. The a matters because expanding (αx + β)(γx + δ) forces the product of the split pieces to be αγβδ = ac — algebra, not convention.

When a = 1 you get away with c because ac = c. The moment a ≠ 1 the shortcut breaks and the method silently fails — telling you a polynomial doesn't factor when it does. The cure costs two seconds: write ac = first, every time. Always use ac.