What this move does
The puzzle above says just one thing:
is a square with a corner missing (an L-shape). The missing corner is a small square of side : half the coefficient of the linear term. Fill it in and you get a perfect square. This move is called completing the square:
- Look at the coefficient of the linear term and take half of it, .
- Add and subtract it again at the same time (or add it to both sides of an equation).
- Write the first three terms as and combine the constants that are left.
When is negative the picture cannot show a “negative length”, but the algebra is exactly the same: .
Predict first
is still missing a corner. How much must you add to make it a perfect square?
Half of the linear coefficient is , so the missing corner is a small square of side and area :
If you chose , you found the right side length but forgot that the corner is an area; if you chose , you forgot to halve first.
Work through one
The most direct use of completing the square is solving equations. The equation says: the area of the L-shape is .
Example: solve x² + 6x = 16
Step 1 / 5- Try any x first
In the picture and the area is , much bigger than . Trying values one by one is too slow, so change tactics: complete the L-shape into a square.
Check : . Correct. The picture can only show the positive root; the algebra finds both.
When the leading coefficient is not 1
Take the leading coefficient out first, and complete the square inside the bracket.
Example: complete the square for 2x² − 12x + 7
Step 1 / 4- Take out the 2
Taking it out of the first two terms is enough; leave the constant outside for now.
The floor of squares
Completing the square has a second use, and it is even neater. The square of any real number is never less than 0: is a floor you cannot break through.
So if several squares add up to , there is only one way it can happen: every one of them is 0. Drag the sliders below and see whether you can make one bar “dig underground” to cancel another.
So one equation can pin down two unknowns at once, as long as you first rewrite it as a sum of squares:
Example: if x² + y² − 6x + 4y + 13 = 0, find x + y
Step 1 / 6- One equation, two unknowns?
Usually an equation like this has infinitely many solutions. But do not give up yet: group the terms together and the terms together.
When it fails
Break it
Counterexample lab: remove one condition
Conclusion: the equation “sum of squares ” has exactly one solution ← still true?
: the constant is shared out exactly between the two squares. Push both bars down to the floor and the total is ; the only solution is , .
Change the constant from to : . After completing the squares there is an extra left over. Push both bars all the way down and see where the total stops.
The total is at least and can never reach : this equation has no real solutions.
The other way round, if the constant is too small (say , that is, ), one term can grow while the other shrinks, and there are infinitely many solutions. Only when the constant is “used up exactly” is a single solution pinned down.
Practice
is a perfect square. Find .
Solve by completing the square. (Separate the two roots with a comma.)
Given , find .
and can be any real numbers. Find the minimum value of .