Visual walkthrough — Polynomial long division and synthetic division

This page builds the synthetic division shortcut from absolute zero. We will not memorise the "bring down, multiply, add" ritual — we will watch it grow out of ordinary long division, one squeeze at a time, until every symbol has earned its place.

Prerequisites we lean on: what a polynomial and its degree are and the master identity from the parent, the topic note. If a word feels new, we define it here before we use it.

Our running example (the same one the parent solves) is: $\underset{\text{dividend }P(x)}{\underbrace{2x^3+3x^2-5x+1}}÷\underset{\text{divisor }D(x)}{\underbrace{(x+2)}}.$

---

Step 1 — Line up the coefficients like place values

WHAT. Write $P(x)$ using only its numbers (coefficients), in descending powers of $x$, with a $0$ wherever a power is missing.

$2x^3+3x^2-5x+1⟶[\underset{x^3}{\underbrace{2}},\underset{x^2}{\underbrace{3}},\underset{x^1}{\underbrace{-5}},\underset{x^0}{\underbrace{1}}]$

• The $2$ lives at power 3 — it is the "thousands column".
• The $1$ lives at power 0 — the "ones column".

WHY. In the number $847$, the $8$ means $800$ only because of its column. Coefficients are identical: the same number means different things depending on its power. Dropping a missing term (like a stray $0x^2$) is exactly the crime of writing $15$ when you meant $1005$ — the columns shift and every later step lands in the wrong place. We keep the columns rigid so the machine never mis-reads a power.

PICTURE. Each coefficient sits in its own labelled column, one per power.

---

Step 2 — Do the first long-division step out loud

WHAT. Long division always asks one question: how many times does the leading term of the divisor go into the leading term of the current dividend?

$\frac{\stackrel{\text{lead of }P}{\overbrace{2x^3}}}{\underset{\text{lead of }D}{\underbrace{x}}}=\underset{\text{first quotient term}}{\underbrace{2x^2}}$

WHY this operation and not another? We want to destroy the highest power one step at a time. The only way to cancel a $2x^3$ by subtracting a multiple of $D(x)=x+2$ is to multiply $D$ by something whose lead is $2x^2$. Division of the lead terms finds that something. That is the entire job of the fraction here: it answers "what quotient term kills the current top?"

Now multiply the whole divisor by that term and subtract: $2x^2\cdot (x+2)=2x^3+4x^2,(2x^3+3x^2)-(2x^3+4x^2)=-x^2.$

• The $2x^3$ cancels exactly — that was the point.
• The new top term is $\underset{=-x^2}{\underbrace{3x^2-4x^2}}$: the divisor's tail ($+2$) leaked into the next column.

PICTURE. Watch the leading $2x^3$ vanish and a fresh $-x^2$ appear.

---

Step 3 — Spot the wasted work

WHAT. Look hard at what actually changed in Step 2. We wrote $2x^3+4x^2$, but the $2x^3$ was guaranteed to cancel — it was born to cancel. The only number that mattered was the $4x^2$ that hit the next column.

$\underset{\text{quotient}}{\underbrace{2}}\times \underset{\text{the }+2\text{ in }D}{\underbrace{2}}=\underset{\text{lands on the }{x}^2\text{ column}}{\underbrace{4}}$

WHY. Half the long-division writing is redundant: the lead term always self-destructs, so copying it out is pointless. If we could agree to never write the doomed lead term again, we'd only ever track two numbers per step: the quotient digit, and the product it pushes into the next column.

PICTURE. The greyed-out $2x^3$ is the wasted ink; the yellow $+4$ is the only thing that moved.

---

Step 4 — Flip a subtraction into an addition

WHAT. In Step 2 we subtracted $+4x^2$. Subtracting is error-prone (sign slips are Mistake 2 in the parent). Watch what happens if we instead subtract $x-c$ by adding $c$.

Our divisor is $x+2=x-(-2)$, so define $c=-2(\text{the value making }x+2=0).$

Then the product that lands in the next column is $\text{quotient digit}\times c=2\times (-2)=-4,$ and we add: $3+(-4)=-1.$

• We got the same $-1$ (i.e. $-x^2$) as Step 2 — but with no subtraction.
• The sign flip from $+4$ to $-4$ is the "subtract"; we absorbed it into $c$.

WHY $c$ and not the divisor's $+2$? Because subtracting $(x+2)$ and adding $(-2)\cdot (\text{quotient})$ produce the identical column change, but addition never trips a sign. This is precisely why synthetic division uses the root $c$, not the visible constant. (Parent Mistake 2: $x+3\Rightarrow c=-3$, never $+3$.)

PICTURE. The subtraction lane and the addition lane give the same result; we keep the addition lane.

---

Step 5 — Roll the whole division into one strip of numbers

WHAT. Repeat the two-number step (multiply by $c$, add the column) across every coefficient. No $x$'s, no leading terms, no subtraction — just:

$\text{bring down}\to \text{multiply by }c\to \text{add}\to \text{repeat}.$

With $c=-2$ and coefficients $[2,3,-5,1]$:

$$\begin{array}{rrrrr}-2& 2& 3& -5& 1\\ & & -4& 2& 6\\ & 2& -1& -3& \boxed{7}\end{array}$$

Tracing each cell so no symbol is unexplained:

• Bring down $2$ — the first quotient coefficient is always the dividend's lead (nothing above it to cancel).
• $2\times (-2)=\underline{-4}$, then $3+(-4)=-1$.
• $-1\times (-2)=\underline{2}$, then $-5+2=-3$.
• $-3\times (-2)=\underline{6}$, then $1+6=\boxed{7}$.

Reading the bottom row: the last entry is the remainder $R=7$. The others are the quotient, one degree lower than $P$ (because dividing by degree-1 $D$ drops the degree by one): $Q(x)=2x^2-1x-3,R=7.$

$\therefore 2x^3+3x^2-5x+1=(x+2)(2x^2-x-3)+7.$

WHY it matches long division exactly. Every bottom-row number is a column-subtraction from long division with the doomed lead term removed and the subtraction re-cast as $+c$. Same arithmetic, less ink.

PICTURE. The full strip, arrows showing multiply-by-$c$ diagonals and add-down columns.

---

Step 6 — The zero-remainder case (why this reveals roots)

WHAT. Try $x^4-3x^2+2÷(x-1)$. Here $c=1$, and the missing $x^3$ and $x$ terms force zeros: coefficients $[1,0,-3,0,2]$.

$$\begin{array}{rrrrrr}1& 1& 0& -3& 0& 2\\ & & 1& 1& -2& -2\\ & 1& 1& -2& -2& \boxed{0}\end{array}$$

• The two $0$s hold the empty columns open — omit them and every later cell shifts wrong (Mistake 1).
• Bottom row: $Q(x)=x^3+x^2-2x-2$, $R=\boxed{0}$.

WHY the zero matters. By the identity $P(x)=D(x)Q(x)+R$, a remainder of $0$ means $P(1)=D(1)Q(1)+0=0$. So $x=1$ is a root — the Remainder & Factor Theorems fall straight out of the last cell. This is why synthetic division is the engine behind the Rational Root Theorem and factoring polynomials.

PICTURE. Same strip; the boxed final $0$ glows to signal "root found — divisor is a factor".

---

Step 7 — The degenerate cases: when synthetic division refuses

WHAT. Two boundary situations you must recognise.

1. Divisor not of the form $x-c$ (e.g. dividing by $2x-6$ or $x^2+1$). The single-number-per-column trick assumes the divisor's lead is $1$ and its degree is $1$. If the lead is $2$, first factor it out: $2x-6=2(x-3)$, divide by $(x-3)$, then divide the quotient by $2$. If the degree is $2+$, synthetic division does not apply — use long division, and stop as soon as $\mathrm{deg}(R)<\mathrm{deg}(D)$ (parent Mistake 4).

2. Dividend degree $<$ divisor degree (e.g. $(3x+1)÷(x-5)$ is fine, but $5÷(x-5)$). If $\mathrm{deg}(P)<\mathrm{deg}(D)$, no term of $D$ fits into $P$: the quotient is $0$ and the whole dividend is the remainder. The strip degenerates to bringing down and adding once.

WHY. The algorithm's promise — $\mathrm{deg}(R)<\mathrm{deg}(D)$ or $R=0$ — is what tells us when to stop and when we're not allowed to start. Respecting it is the difference between a correct answer and Mistakes 1 & 4.

PICTURE. A decision panel: is the divisor $x-c$? is $\mathrm{deg}P\ge \mathrm{deg}D$? Only the green path runs the strip.

---

The one-picture summary

Every idea on this page, compressed: long division's redundant lead term is deleted, its subtraction is re-cast as "add $c$", and what survives is a single strip of multiply-diagonals and add-columns whose final cell is the remainder.

Recall Feynman retelling — say it in plain words

Long division of polynomials wastes effort: every time you divide, the top term is designed to cancel, so writing it out again is pointless. Throw it away. What's left to track is just one number per column. Then notice that "subtract the divisor" secretly equals "add $c$", where $c$ is the value that makes $x-c$ zero — so we replace every subtraction with an addition and never slip a sign. Slide those two moves across the row of coefficients: bring the first one down, multiply by $c$ and drop it into the next column, add, repeat. The numbers that fall out the bottom are the answer (the quotient), and the very last one is what's left over (the remainder). If that last number is $0$, then $x=c$ divides in cleanly — $c$ is a root — which is exactly how we hunt for factors. Just keep the empty powers marked with $0$, and don't try this when the divisor is anything fancier than $x-c$.

---

Related next step: once you can force remainders to zero, splitting a fraction becomes easy — see partial fractions.