To write a quadratic function in vertex form by completing the square, you rewrite y = ax² + bx + c as y = a(x − h)² + k, where the vertex is (h, k) and h = −b/(2a). The reliable route is to factor out a first, then build a perfect square inside the parentheses and rebalance the constant outside. If you skip the factoring step when a ≠ 1, the vertex comes out wrong by a factor of a. The method works for every quadratic, including ones with fractional or negative coefficients, and you can always check the result by expanding back to standard form.
Why Vertex Form Is Worth the Rewrite
y = a(x − h)² + k tells you the vertex coordinates directly: (h, k). Standard form y = ax² + bx + c tells you the y-intercept (0, c) and nothing about the turning point without extra work. That difference matters when you're asked for the maximum height of a projectile, the minimum cost in a cost function, or the axis of symmetry of a parabola on a graph.
The leading coefficient a carries over unchanged between the two forms. That's the single most useful fact for self-checking: if your standard form starts with 2x², your vertex form must still start with 2. A finished answer beginning with a different leading coefficient means an arithmetic slip somewhere in the middle.
Completing the square also produces the quadratic formula as a by-product when you apply it to the general case ax² + bx + c = 0. Textbooks such as OpenStax's Elementary Algebra 2e teach the derivation this way precisely because the rewriting skill and the formula share one mechanical core.
One boundary worth stating: completing the square works over the real numbers for every quadratic, but it does not always produce integer coordinates. A large share of textbook and exam problems yield fractional h and k values, so a non-integer answer is normal, not a signal of error.
The Standard Procedure in Five Moves
Here is the sequence in the order you should actually perform it, using y = 2x² − 12x + 11 as the running example.
1: Write the function with the x² and x terms grouped. y = (2x² − 12x) + 11. Leave the constant outside the parentheses. 2: Factor out a from the grouped terms only. y = 2(x² − 6x) + 11. Note that 11 is untouched. 3: Take half of the new x-coefficient and square it. Half of −6 is −3; (−3)² = 9. 4: Add and subtract that number inside the parentheses. y = 2(x² − 6x + 9 − 9) + 11. Adding and subtracting the same value inside doesn't change the expression. 5: Fold the perfect square and move the subtracted term out. y = 2(x − 3)² − 18 + 11 = 2(x − 3)² − 7.
If any step produces a sign error, the vertex lands on the wrong side of the y-axis. The most common failure case is forgetting to multiply the subtracted term by a before moving it outside; in the example above, that produces −9 + 11 = 2 instead of −18 + 11 = −7, and the vertex shifts from (3, −7) to (3, 2). Re-expanding at the end catches this in about fifteen seconds.
Worked Example With Whole-Number Coefficients
Take y = x² + 8x + 5. Because a = 1, no factoring step is needed and the procedure shortens.
Group and prepare: y = (x² + 8x) + 5. Half of 8 is 4, and 4² = 16. Add and subtract inside: y = (x² + 8x + 16 − 16) + 5. The perfect square collapses: y = (x + 4)² − 16 + 5, so y = (x + 4)² − 11.
Reading the result: the vertex is (−4, −11). Watch the sign convention here, since (x + 4)² is the same as (x − (−4))², so h = −4, not +4. This sign flip is the second most frequent student error after the factoring mistake.
Verification: expand (x + 4)² − 11 = x² + 8x + 16 − 11 = x² + 8x + 5, which matches the original. Doing this expansion in writing rather than in your head is worth the extra line, because the arithmetic is where errors hide.
A quick graph check follows from the same result. The parabola opens upward because a = 1 > 0, has its lowest point at (−4, −11), and crosses the y-axis at (0, 5), which is 16 units above the vertex.
The a ≠ 1 Case and the Sign That Trips People
When the leading coefficient isn't 1, the subtracted square must be multiplied by a before it leaves the parentheses. That single multiplication is where most errors originate.
Run y = 3x² + 6x + 1. Factor: y = 3(x² + 2x) + 1. Half of 2 is 1, and 1² = 1. Add and subtract: y = 3(x² + 2x + 1 − 1) + 1. Fold: y = 3(x + 1)² − 3(1) + 1 = 3(x + 1)² − 2. Vertex: (−1, −2).
Now the failure case in full. A student who writes y = 3(x + 1)² − 1 + 1 = 3(x + 1)² gets a vertex of (−1, 0), which is wrong by 2 units vertically. The error is invisible without a check, because the expression still looks plausible.
Confirm with the vertex formula h = −b/(2a) = −6/(2·3) = −1, which agrees. Substitute x = −1 into the original: 3(1) − 6 + 1 = −2, matching k. Both checks are independent of the completing-the-square work, which is why they catch different mistakes.
For a negative leading coefficient, factor out the negative along with the number. In y = −x² + 4x − 1, factor −1: y = −(x² − 4x) − 1, then complete: y = −(x² − 4x + 4 − 4) − 1 = −(x − 2)² + 4 − 1 = −(x − 2)² + 3, vertex (2, 3), opening downward.
Fractional Coefficients Without Panic
Fractions don't change the method; they change the arithmetic volume. Take y = x² − 3x + 1. Half of −3 is −3/2, and (−3/2)² = 9/4.
Add and subtract: y = (x² − 3x + 9/4 − 9/4) + 1 = (x − 3/2)² − 9/4 + 4/4 = (x − 3/2)² − 5/4.
Vertex: (3/2, −5/4) = (1.5, −1.25). Formula check: h = −(−3)/(2·1) = 3/2, and substituting gives 9/4 − 9/2 + 1 = 9/4 − 18/4 + 4/4 = −5/4.
The arithmetic trap here is adding a whole number to a fraction with denominator 4. Convert 1 to 4/4 first, then combine. Writing the conversion explicitly prevents the common answer −13/4, which appears when someone adds numerators across different denominators.
If the coefficient of x is already a fraction, say y = x² + (2/3)x + 5, half of 2/3 is 1/3 and its square is 1/9, giving y = (x + 1/3)² − 1/9 + 45/9 = (x + 1/3)² + 44/9. The pattern is identical; only the denominators grow.
Converting Back to Standard Form as a Check
Expanding is the fastest verification available because it uses a different operation than the forward method. If you complete the square forward and expand backward and land on the original expression, the probability of a compensating error is very low.
The expansion rule is mechanical: a(x − h)² + k = a(x² − 2hx + h²) + k = ax² − 2ahx + ah² + k. Compare coefficients with ax² + bx + c to get b = −2ah and c = ah² + k. Both equations must hold.
Apply that to the earlier result y = 2(x − 3)² − 7. Expansion gives 2(x² − 6x + 9) − 7 = 2x² − 12x + 18 − 7 = 2x² − 12x + 11, which is the original. The coefficient comparison confirms h = 3 from −12 = −2·2·3 and c = 2·9 − 7 = 11.
If your expansion misses by a constant but the x-coefficient matches, the error is in the k value only, usually a dropped multiplication by a. If the x-coefficient is wrong, the error is in h, which traces back to taking half the coefficient incorrectly.
Which Form Answers Which Question
Choosing between forms isn't a style preference; each one exposes different information with zero extra computation. Use this table to decide what to produce when a problem asks a specific question.
| Question asked | Form to use | Where the answer sits | Worked value for y = 2x² − 12x + 11 |
|---|---|---|---|
| Vertex coordinates | Vertex form | (h, k) directly | (3, −7) |
| Axis of symmetry | Vertex form | x = h | x = 3 |
| Maximum or minimum value | Vertex form | k, with direction from sign of a | Minimum −7 |
| y-intercept | Standard form | (0, c) | (0, 11) |
| Number of x-intercepts | Discriminant from standard form | b² − 4ac compared to 0 | 144 − 88 = 56, two intercepts |
| Roots or zeros | Factored form or quadratic formula | Solve a(x − h)² + k = 0 | x = 3 ± √(7/2) |
The last row is worth dwelling on, because vertex form makes the roots easy: set 2(x − 3)² − 7 = 0, so (x − 3)² = 7/2 and x = 3 ± √3.5 ≈ 3 ± 1.871, giving roughly 4.87 and 1.13. Standard form requires the quadratic formula for the same answer with more substitution.
A narrow but real cost: producing vertex form adds roughly 30 to 60 seconds of arithmetic compared with reading c straight off standard form. When a problem asks only for the y-intercept, converting is wasted effort.
Mistakes That Survive a Careless Check
Four errors account for most wrong answers, and each has a distinguishing signature. The factoring omission shifts k by a multiple of a, as shown earlier. The sign flip on h produces a vertex mirrored across the y-axis while the k value stays correct. Dropping the subtraction inside the parentheses makes the parabola's vertex sit at the origin of the shifted term. And combining fractions without a common denominator corrupts k alone.
A subtler edge case: functions with no real roots, such as y = x² + 4x + 9. Completing the square gives y = (x + 2)² + 5. The vertex is (−2, 5) and the parabola never touches the x-axis, since the minimum value 5 is positive. Students sometimes assume completing the square failed here because no x-intercepts appear; nothing failed, and this result is exactly what vertex form is for.
The rarer degenerate case is a = 0. Then the function is linear, not quadratic, and the method doesn't apply. A problem presenting something like y = 0x² + 5x + 1 is hiding a line.
Textbook problem sets frequently include one of these traps. Recognizing them before starting saves the whole calculation.
What the Formula Check Reveals That Expansion Misses
Expansion confirms the algebra is consistent; it cannot confirm the vertex is actually an extremum. The formula route h = −b/(2a) does, because it derives from the derivative condition 2ax + b = 0, which states where the slope is zero.
Practical protocol: compute h both ways. If completing the square gives (3, −7) and the formula gives h = 12/4 = 3, the horizontal coordinate is confirmed by two methods. Then compute k by substitution into the original function: 2(9) − 36 + 11 = −7. Three independent checks, each about ten seconds.
For fractional cases, do the formula check in fractions rather than decimals to avoid rounding masks. In y = x² − 3x + 1, h = 3/2 exactly; writing 1.5 is fine, but writing 1.49 and accepting it invites a hidden error.
One limitation of the formula check: it gives h and k but not the squared factor's sign convention. If you write (x + 3)² when the correct term is (x − 3)², the formula check on b catches it because the signs disagree.
Reading the Graph Straight Off the Finished Form
Vertex form translates into graph features without plotting points. In y = −2(x + 1)² + 8, the vertex is (−1, 8), the parabola opens downward because a = −2, and the axis of symmetry is the vertical line x = −1.
The x-intercepts come from setting y = 0: −2(x + 1)² + 8 = 0, so (x + 1)² = 4, x + 1 = ±2, x = 1 or x = −3. The y-intercept comes from x = 0: y = −2(1) + 8 = 6.
Compare that to standard form y = −2x² − 4x + 6, where the vertex requires either the formula or a conversion. The same parabola needs about four additional lines of work to describe from standard form.
Intervals of increase and decrease drop out too. For a < 0, the function increases on x < h and decreases on x > h. Here it increases on x < −1 and decreases on x > −1. No calculus, no test points.
Sensory check for anyone sketching: the vertex sits at the peak, the arms drop away symmetrically, and the curve crosses the y-axis at 6, well below the peak of 8. That picture takes about thirty seconds once the vertex form is in hand.
A Complete Rewrite From Start to Verified Answer
Put the whole method on one problem with awkward numbers: y = 5x² + 20x − 1.
Factor: y = 5(x² + 4x) − 1. Half of 4 is 2, square is 4. Add and subtract: y = 5(x² + 4x + 4 − 4) − 1. Fold: y = 5(x + 2)² − 20 − 1 = 5(x + 2)² − 21. Vertex: (−2, −21).
Formula check: h = −20/(2·5) = −2. Substitution: 5(4) − 40 − 1 = 20 − 40 − 1 = −21. Expansion: 5(x² + 4x + 4) − 21 = 5x² + 20x + 20 − 21 = 5x² + 20x − 1. All three agree.
The dangerous error in this problem is the −20. Someone who subtracts 4 rather than 20 gets −5, and the vertex reads (−2, −5). Both numbers appear in the working, so the result looks reasonable; only the check catches it.
A reporting note: write the final answer with the vertex on its own line, since exam rubrics typically award the vertex point separately from the correct rewritten expression.
FAQ
What is the vertex of y = x² − 10x + 22?
The vertex is (5, −3). Completing the square gives y = (x − 5)² − 3, and the formula check h = 10/2 = 5 agrees. Substitute x = 5 into the original: 25 − 50 + 22 = −3. The parabola opens upward, so −3 is a minimum.
Does completing the square always produce a perfect square with integer terms?
No. When the x-coefficient is odd, half of it is a fraction whose square has a denominator of 4, as in y = x² − 3x + 1 giving (x − 3/2)² − 5/4. Integer results occur only when b is even after factoring out a, or when the fraction simplifies.
Why do I get the wrong vertex when I forget to multiply by a?
The expression inside the parentheses is multiplied by a, so anything subtracted there is also multiplied by a. Leaving it unscaled shifts k by (a − 1) times the squared half-coefficient. For y = 3x² + 6x + 1, that error gives k = 0 instead of −2.
Can I use the quadratic formula instead of completing the square to find the vertex?
The formula h = −b/(2a) gives the same horizontal coordinate and comes from the same derivation. It's faster if you only need the vertex. Completing the square also yields the full rewritten expression, which the formula doesn't provide, so for the specific task of writing vertex form, completing the square is the direct route.
What if the quadratic has no real x-intercepts?
Completing the square still works. For y = x² + 4x + 9, the result is (x + 2)² + 5 with vertex (−2, 5). The function never crosses the x-axis because its minimum is positive. No real roots is a feature of the function, not a failure of the method.
Is vertex form the same as the quadratic formula?
No. Vertex form is an equivalent way to write the function that displays the vertex. The quadratic formula is a solving tool derived by completing the square on ax² + bx + c = 0. One outputs an expression; the other outputs the roots.
How long does the rewrite take to check by expansion?
Expansion takes roughly fifteen seconds for a simple case and about thirty for one with a fractional constant. It catches sign errors and missed multiplications that a formula check on h alone misses, because it tests the full expression including the constant term.
Can I complete the square when a is negative?
Yes. Factor the negative coefficient out along with the number, so y = −x² + 4x − 1 becomes y = −(x² − 4x) − 1, then complete inside as usual. The leading coefficient −1 stays in front, and the parabola opens downward from the resulting vertex.