Factoring only finds the vertex quickly when the roots are nice integers, and many quadratics do not factor at all. Completing the square works on every quadratic, no matter how awkward the numbers are, and it converts standard form directly into vertex form. That matters because vertex form answers the questions people actually ask about quadratic models: what is the greatest height of a ball, the lowest cost, the largest area or the maximum revenue, and at what input does it happen? Completing the square is also the process that derives the quadratic formula in the next lesson, so understanding it here means the formula will never feel like magic. The whole method rests on one pattern, the perfect square trinomial, and one honest trick: adding zero in a clever form.
The idea: make a perfect square trinomial
Expand any squared binomial and look at the pattern: . The middle coefficient is and the constant is , so the constant is always the square of half the -coefficient. For example factors as because . The picture explains why. Build from an -by- square plus two -by- strips placed along two sides. The shape is almost a big square of side , but one corner is missing: a -by- piece with area . "Completing the square" literally means filling in that missing corner. Since we cannot change the value of the expression, we add the and immediately subtract it again, which is the same as adding zero.
Case 1: leading coefficient a = 1
- Take the coefficient of , halve it, and square the result: . For : half of is , and .
- Add and subtract that number right after the -term: .
- Group the first three terms as a perfect square binomial, using half of (with its sign) inside the bracket: .
- Simplify the leftover constants: , so the vertex is .
Two details make this reliable. First, the number inside the bracket is always half of with the same sign as : becomes , and becomes . Second, odd values of simply produce fractions, and the method still works. For , half of is and its square is , so The vertex is , or . Keep fractions exact until the last line; rounding early spoils the vertex.
Case 2: leading coefficient a ≠ 1
The perfect-square pattern only works when has coefficient . So when , first factor out of the and terms only, leaving the constant term outside and untouched. Complete the square inside the bracket as before, then distribute back through before simplifying. The coefficient inside may become a fraction, which is fine: becomes ; half of is , squared is , so When is negative, factor out the negative too: becomes , and every sign inside the bracket flips.
Reading the maximum or minimum
Once in vertex form , the vertex is . If , is the minimum value of the function; if , is the maximum value — because a squared term is always , so only ever adds to (when ) or subtracts from it (when ), and it equals zero only at . This is why completing the square is the standard tool for optimization. The completed form also tells you about the zeros. Solving gives , which has two solutions when , one when , and none when (a square cannot be negative). For : , so .
| Standard form | Completed square | Vertex | Max or min |
|---|---|---|---|
| min | |||
| min | |||
| min | |||
| min | |||
| max |