← mapFunctions & Growth

Quadratic Functions & Parabolas

⚗ Dr. Möbius, from the lab

A parabola is not just "a U-shape." It is the graph of the most fundamental nonlinear relationship in algebra, and it encodes a maximum or minimum that has been optimizing everything from cannonball trajectories to fencing budgets since before calculus existed. You already derived the quadratic formula by completing the square — now you're going to see that completing the square was secretly building the parabola's soul the whole damn time, and every textbook that skips this connection should be confiscated and burned in the reactor.

THE BIG IDEA

Every quadratic function has a vertex that is its single most important point, and completing the square reveals that vertex directly while connecting the formula to the geometry.

From equation to function

In the Algebra stratum, you solved quadratic equations — you found the xx-values where ax2+bx+c=0ax^2 + bx + c = 0. Now we ask a bigger question: what does the entire function f(x)=ax2+bx+cf(x) = ax^2 + bx + c look like? What shape is the graph, and what does every coefficient do?

The answer is: the graph is a parabola, and the coefficient aa is its whole personality. One number. That's all it takes to tell me whether this thing is a cup or a cap, a smile or a frown.

f(x)=ax2+bx+c,a0.f(x) = ax^2 + bx + c, \qquad a \ne 0.

  • If a>0a > 0: the parabola opens upward (cup shape, \cup). It has a minimum.
  • If a<0a < 0: the parabola opens downward (cap shape, \cap). It has a maximum.
  • Larger a|a|: narrower parabola. Smaller a|a|: wider.

The parabola is symmetric — there's a vertical line through its lowest (or highest) point that is a mirror axis. That point is the vertex, and it's the single most important piece of information the parabola carries. Everything else is decoration.

Use the lab below to see how aa, bb, cc each transform the parabola:

parabola lab — y = ax² + bx + c
vertex (0, 0)b²−4ac = 0roots x = 0

Vertex form: the completing-the-square payoff

You completed the square to derive the quadratic formula — I know, I was there, the lab smelled like chalk and regret. Now watch the same move produce the vertex form.

Start from f(x)=ax2+bx+cf(x) = ax^2 + bx + c. Factor aa from the first two terms:

f(x)=a ⁣(x2+bax)+c.f(x) = a\!\left(x^2 + \frac{b}{a}x\right) + c.

Complete the square inside the parentheses. The magic number to add (and subtract) is (b2a)2\left(\frac{b}{2a}\right)^2:

f(x)=a ⁣(x2+bax+b24a2b24a2)+cf(x) = a\!\left(x^2 + \frac{b}{a}x + \frac{b^2}{4a^2} - \frac{b^2}{4a^2}\right) + c =a ⁣(x+b2a)2b24a+c.= a\!\left(x + \frac{b}{2a}\right)^2 - \frac{b^2}{4a} + c.

Define h=b2ah = -\dfrac{b}{2a} and k=cb24ak = c - \dfrac{b^2}{4a}. Then:

f(x)=a(xh)2+k.\boxed{f(x) = a(x - h)^2 + k.}

This is vertex form. Why? Because (xh)20(x - h)^2 \ge 0 always, and it equals 00 exactly when x=hx = h. So:

  • The minimum (if a>0a > 0) or maximum (if a<0a < 0) output is kk.
  • It occurs at x=hx = h.
  • The vertex is (h,k)(h, k).

The vertex IS the function's personality in a single point. Everything else follows from it and aa. If you remember nothing else from this lesson — though I expect you to remember everything — remember the vertex form and how to find it.

The axis of symmetry

The axis of symmetry is the vertical line x=h=b2ax = h = -\dfrac{b}{2a}.

Why? The quadratic in vertex form, a(xh)2+ka(x-h)^2 + k, depends on (xh)2(x-h)^2 — a squared quantity. Squaring is indifferent to sign: (h+d)2=(hd)2(h + d)^2 = (h - d)^2. So f(h+d)=f(hd)f(h + d) = f(h - d) for any dd. The function takes equal values at equal distances left and right of hh.

Consequence: if the parabola has two real roots r1r_1 and r2r_2, they are symmetric around x=hx = h. Their average is hh:

r1+r22=h=b2a.\frac{r_1 + r_2}{2} = h = -\frac{b}{2a}.

(This is also Vieta's formula for the sum of roots: r1+r2=b/ar_1 + r_2 = -b/a, so their average is b/2a-b/2a.)

The discriminant, drawn

You know the discriminant Δ=b24ac\Delta = b^2 - 4ac from the Algebra stratum. Now see it geometrically:

  • Δ>0\Delta > 0: two distinct real roots. Parabola crosses the xx-axis in two places.
  • Δ=0\Delta = 0: exactly one root (repeated). Parabola is tangent to the xx-axis — the vertex sits exactly on it.
  • Δ<0\Delta < 0: no real roots. Parabola misses the xx-axis entirely — vertex is strictly above (if a>0a > 0) or below (if a<0a < 0).

The discriminant tells you how many intersections the parabola has with the xx-axis. That's it. Three answers; three pictures. Not six. Not twelve. Three. This should be the fastest check you do in the whole problem.

Max/min word problems: the first taste of optimization

The vertex gives the maximum or minimum output. Whenever a real-world problem asks "what is the greatest profit / shortest time / maximum area?" and the model is a quadratic, your job is exactly: find the damn vertex. Don't integrate. Don't differentiate. Just find the vertex — we built the machinery for exactly this.

Standard setup: fencing. You have 200200 meters of fencing to enclose a rectangular plot against a wall (so only three sides need fencing). If the side perpendicular to the wall has length xx, the side parallel has length 2002x200 - 2x, and the area is:

A(x)=x(2002x)=2x2+200x.A(x) = x(200 - 2x) = -2x^2 + 200x.

This is a downward parabola (a=2<0a = -2 < 0), so it has a maximum. The vertex is at:

x=b2a=2002(2)=50.x = -\frac{b}{2a} = -\frac{200}{2(-2)} = 50.

Maximum area: A(50)=50(200100)=50100=5000A(50) = 50 \cdot (200 - 100) = 50 \cdot 100 = 5000 square meters.

This exact move — write the quantity as a quadratic, find the vertex — is the entire engine of optimization before calculus enters. File it. You'll need it more times than you expect.

🔬 SPECIMENS (worked examples)

Worked example 1 — converting to vertex form and reading the vertex

Write f(x)=2x28x+3f(x) = 2x^2 - 8x + 3 in vertex form and identify the vertex.

Factor out a=2a = 2 from the xx-terms: f(x)=2(x24x)+3.f(x) = 2(x^2 - 4x) + 3.

Complete the square: half of 4-4 is 2-2; (2)2=4(-2)^2 = 4. Add and subtract 44 inside: f(x)=2(x24x+44)+3=2((x2)24)+3.f(x) = 2(x^2 - 4x + 4 - 4) + 3 = 2\bigl((x-2)^2 - 4\bigr) + 3.

Distribute the 22: f(x)=2(x2)28+3=2(x2)25.f(x) = 2(x-2)^2 - 8 + 3 = 2(x-2)^2 - 5.

Vertex form: f(x)=2(x2)25f(x) = 2(x-2)^2 - 5.

Vertex: (h,k)=(2,5)(h, k) = (2, -5).

Check: f(2)=2(0)25=5f(2) = 2(0)^2 - 5 = -5. Confirmed. Since a=2>0a = 2 > 0, this is the minimum value of ff.

Worked example 2 — discriminant and sketching the parabola

For g(x)=x2+4x5g(x) = -x^2 + 4x - 5, find the vertex, determine how many xx-intercepts the parabola has, and sketch its rough shape.

Here a=1a = -1, b=4b = 4, c=5c = -5.

Vertex. Axis of symmetry: x=b2a=42(1)=2x = -\frac{b}{2a} = -\frac{4}{2(-1)} = 2. Vertex yy-value: g(2)=(4)+85=1g(2) = -(4) + 8 - 5 = -1. Vertex: (2,1)(2, -1).

Discriminant. Δ=b24ac=164(1)(5)=1620=4<0\Delta = b^2 - 4ac = 16 - 4(-1)(-5) = 16 - 20 = -4 < 0.

Since Δ<0\Delta < 0: no real xx-intercepts. The parabola never touches the xx-axis.

Shape. Since a=1<0a = -1 < 0: opens downward. Vertex at (2,1)(2, -1) is the maximum. The entire parabola lives below y=1y = -1, so it never crosses y=0y = 0. The sketch is a downward-opening cap with its peak at (2,1)(2, -1), floating entirely below the xx-axis.

Worked example 3 — fencing optimization (vertex as the punchline)

A farmer has 120 meters of fencing to enclose a rectangle. What dimensions maximize the enclosed area, and what is that maximum area?

Let one side have length xx. Since the perimeter is 2x+2y=1202x + 2y = 120, we get y=60xy = 60 - x.

Area: A(x)=x(60x)=x2+60xA(x) = x \cdot (60 - x) = -x^2 + 60x.

This is a downward parabola (a=1<0a = -1 < 0), so it has a maximum.

Vertex: x=b2a=602(1)=30x = -\frac{b}{2a} = -\frac{60}{2(-1)} = 30.

So y=6030=30y = 60 - 30 = 30 as well. The optimal rectangle is a square, 30×3030 \times 30.

Maximum area: A(30)=3030=900A(30) = 30 \cdot 30 = 900 square meters.

The trap everyone falls into: using calculus instincts before reading aa. You don't need calculus — the vertex of a downward parabola is automatically the maximum, and you can find it algebraically. The quadratic's vertex formula IS the pre-calculus version of "set the derivative to zero."

☠ KNOWN HAZARDS

  • Misreading vertex form. f(x)=(x3)2+1f(x) = (x - 3)^2 + 1 has vertex (3,1)(3, 1), not (3,1)(-3, 1). The hh in (xh)2(x - h)^2 is the xx-coordinate of the vertex; the minus sign is baked in. Every year, approximately half the students on Earth get this backwards. Do not be that half.

  • Forgetting to subtract what you added inside the parentheses. When completing the square, if you add (b2a)2\left(\frac{b}{2a}\right)^2 inside, you have added a(b2a)2a \cdot \left(\frac{b}{2a}\right)^2 to the function (because the aa out front multiplies everything inside). Compensate by subtracting that same quantity. The bookkeeping has to be exactly right — "close enough" is bullshit in algebra.

  • Treating Δ<0\Delta < 0 as "no answer." It means no REAL roots — the parabola lives entirely above or below the xx-axis. The function is still perfectly defined on all of R\mathbb{R}; it just never equals zero. Complex roots exist but aren't covered here. "No real roots" is an answer, not a failure.

  • Optimizing without checking aa. Before celebrating your vertex as a maximum, verify a<0a < 0. If a>0a > 0, the vertex is a minimum, and the function has no finite maximum at all. Confusing a minimum for a maximum on an exam is the kind of mistake that makes me want to flip the beakers.

TL;DR

  • The graph of f(x)=ax2+bx+cf(x) = ax^2 + bx + c is a parabola: a>0a > 0 opens up (minimum), a<0a < 0 opens down (maximum). Larger a|a| makes it narrower.

  • Vertex form f(x)=a(xh)2+kf(x) = a(x-h)^2 + k is produced by completing the square; the vertex is (h,k)= ⁣(b2a,  cb24a)(h, k) = \!\left(-\frac{b}{2a},\; c - \frac{b^2}{4a}\right).

  • The axis of symmetry x=b2ax = -\frac{b}{2a} is the mirror line; roots (when real) are symmetric around it.

  • The discriminant Δ=b24ac\Delta = b^2 - 4ac tells you how many times the parabola crosses the xx-axis: two (Δ>0\Delta > 0), one (Δ=0\Delta = 0), none (Δ<0\Delta < 0).

  • Max/min word problems: model the quantity as a quadratic, then find the vertex — that's optimization before calculus.

unlocks