A quadratic is a parabola in disguise: completing the square, the formula, and factoring are three views of the same structure. Polynomials extend the pattern.
Core ~14 minDeep study ~32 minWhy does completing the square produce the quadratic formula?What does the discriminant decide geometrically?
Merchants and surveyors met problems that linear rules could not finish: given an area and how much length exceeds width, find the sides; given a square and something added beside it, restore a pure square. Those are quadratic stories - unknown times unknown appears.
MacTutor's history of quadratic equations warns against a slogan that "the Babylonians invented the quadratic formula." Old Babylonian tablets (second millennium BCE) show algorithmic procedures, essentially completing the square, for problems we would write as quadratics. They had no modern notion of "equation," and answers were positive lengths. Euclid later gave geometric square-completion. Hindu mathematicians (notably Brahmagupta) admitted negatives more freely. al-Khwārizmī (c. 9th century) classified types and proved solutions by geometry because zero and negative coefficients were not yet part of his toolkit.
For a modern rereader, ax2+bx+c=0 (with a=0) is one family. Completing the square, the quadratic formula, and factoring are three cameras on the same object. The graph is a parabola.
First principles
Fix area, vary one side: the other side is A/x. Plotting that height against x draws a hyperbola; related quadratic shapes appear when you expand (x)(s-x).
A quadratic function is f(x)=ax2+bx+c with a=0. Its graph is a parabola opening up (a>0) or down (a<0).
The vertex sits at x=−b/(2a). Completing the square rewrites
f(x)=a(x+2ab)2+(c−4ab2)
so the shift and the minimum/maximum height are visible.
The quadratic formula for ax2+bx+c=0 is
x=2a−b±b2−4ac
The discriminantΔ=b2−4ac decides geometry:
Δ>0: two distinct real roots (two x-intercepts)
Δ=0: one repeated real root (vertex on the axis)
Δ<0: no real roots (parabola misses the x-axis); complex roots exist in C
Factoring over the reals works when roots are nice: if f(x)=a(x−r)(x−s), roots are r and s.
A polynomial of degree n is anxn+⋯+a0 with an=0. Quadratics are degree 2. Higher degrees bend more; the fundamental theorem of algebra (later) says degree n has n roots in C, counting multiplicity.
Worked intuition
Completing the square by hand
Solve x2+6x+5=0.
x2+6x=−5. Half of 6 is 3; square is 9. Add 9: (x+3)2=4. So x+3=±2, hence x=−1 or x=−5.
Formula as packaged completion
Apply the formula to 2x2−4x−6=0:
x=44±16+48=44±8
Roots 3 and −1. Same answers as factoring 2(x−3)(x+1)=0.
Reading the parabola
For y=x2−4x+3, vertex at x=2, y=−1. Roots at 1 and 3. Axis of symmetry x=2. Sketch a few points and the U-shape from the graphs lesson returns with a name.
Polynomial preview
x3−x=x(x−1)(x+1) shows factoring still organizes higher degree when you can find roots. Not every cubic factors nicely over the integers; that is why later tools exist.
Common confusions
Forgetting a=0. If a=0, the equation is linear, not quadratic.
Dropping the ±. Completing the square always opens two branches unless the square root is zero.
Thinking Babylonians had modern formulas. They had algorithms for positive length problems. Respect both the achievement and the conceptual gap.
Confusing vertex formula with roots.−b/(2a) is the axis; roots need the discriminant.
Assuming every quadratic factors with integers. The formula always works (in C); nice integer factoring is a gift, not a right.
History and stories
MacTutor's quadratic history is the honest spine: Babylonian completing-the-square algorithms for area problems; Euclid's geometry; Brahmagupta's more algebraic treatment with negatives; al-Khwārizmī's classified cases and geometric proofs; later European symbolic algebra packing cases into one formula.
al-Khwārizmī's ethic is pedagogy: show a method a stranger can repeat, then justify it with a diagram. That is why his name echoes in "algebra" and "algorithm."
Real world
Projectiles (idealized): height versus time is quadratic under constant gravity.
Area fencing problems: maximize area for fixed perimeter leads to a parabola in the model.
Optics and mirrors: parabolic reflectors focus parallel rays (geometry of the curve).
Business: revenue models with price and quantity sometimes produce quadratic profit.
Open when you want the machine-learning connection. Skip freely.
Field notes, extra examples, and glossary live here when you want more than the core path.
Carry this forward
Parabola, completing the square, formula, and factoring are one family.
Discriminant is geometry in a number.
Polynomials continue the degree ladder.
Next: exponentials and logarithms, where growth stops being polynomial.
Read more
Go deeper when curiosity hits - videos, essays, and primary trails.