I’m struggling to complete the square reliably once the quadratic has a leading coefficient that isn’t 1, or when the linear coefficient is odd. Do I always need to factor the leading coefficient out of the x^2 and x terms first, even if it’s negative, and why is that step logically required rather than just a shortcut? When the linear coefficient is odd, fractions show up-am I supposed to accept the fractions early, or is there a neat way to delay them without making mistakes? In an equation, what’s the most dependable rule to keep both sides balanced-should I add the same amount to both sides immediately, or is it okay to adjust inside parentheses and then compensate later? How can I quickly check that the binomial square I end up with is correct without fully expanding everything again? Lastly, what is the cleanest path from completing the square to the vertex form in these messier cases, and how do I read the vertex off confidently?
Ready to make maths more enjoyable, accessible, and fun? Join a friendly community where you can explore puzzles, ask questions, track your progress, and learn at your own pace.
By becoming a member, you unlock:
3 Responses
Short version: start with ax^2 + bx + c, and always pull a out of the x^2 and x terms first: a(x^2 + (b/a)x) + c. That’s not a “nice-to-have,” it’s the logic-completing the square uses (x + p)^2 = x^2 + 2px + p^2, which only matches when the x^2 coefficient is 1; otherwise you’d be trying to square (√a x + …), which is messier. Inside the bracket, take half of the x-coefficient: p = b/(2a), then add and subtract p^2 inside, or more cleanly write a[(x + p)^2 − p^2] + c = a(x + p)^2 + c − a·p^2, so vertex form pops out as a(x + b/(2a))^2 + (c − b^2/(4a)); vertex is at (−b/(2a), c − b^2/(4a)). Fractions when b is odd? Don’t fight them; they’re inevitable for vertex form. If you really hate them while solving an equation, multiply the whole thing by 4a and complete the square in 2ax: (2ax + b)^2 = b^2 − 4ac-no fractions until the end. Balancing rule I trust: whatever you add inside parentheses gets multiplied by a, so either (a) add it inside and immediately subtract the same a·p^2 outside, or (b) if it’s an equation, add a·p^2 to both sides-just be consistent with the factor a. Quick sniff test of your binomial: if you’ve got a(x + h)^2 + k, then 2ah should equal b (with the right sign), and h should equal −b/(2a); that catches most slips without fully expanding. Example: 3x^2 + 5x − 2 = 3[x^2 + (5/3)x] − 2 = 3[(x + 5/6)^2 − (25/36)] − 2 = 3(x + 5/6)^2 − 25/12 − 2 = 3(x + 5/6)^2 − 49/12, so vertex is (−5/6, −49/12). When I finally stopped dodging the fractions and just did the p = b/(2a) step cleanly, my error rate tanked-my old “delay the fractions” trick was exactly how I kept losing the factor of a.
The north star is this: any perfect square that starts with ax^2 must look like a(x + h)^2, because expanding gives ax^2 + 2ah x + a h^2. That’s why we factor a from the x^2 and x terms first (even if a is negative): it resets the leading term inside the parentheses to x^2 so the “half-the-coefficient” trick applies cleanly. After factoring, the x-coefficient inside is b/a, so h must be b/(2a). If b is odd, h is fractional-totally fine. For a slick but optional way to dodge fractions when solving an equation, you can multiply the whole equation by 4a first; then 4a^2 x^2 + 4ab x turns into a tidy (2ax + b)^2 after adding b^2. For vertex form, embracing the fraction is simplest: y = a(x − h)^2 + k with h = −b/(2a) and k = c − b^2/(4a).
When balancing an equation, the safest habit is: if you add t inside a( … ), you’ve really added a·t to that side-so add a·t to the other side immediately. It’s okay to compensate later, but that’s when gremlins steal your factors of a. Quick correctness check of your binomial: after you’ve factored a, your square should be (x + h)^2 with h = b/(2a). Double h should match the inside x-coefficient, and the “added constant” outside should be a·h^2. Reading the vertex is then a breeze: y = a(x − h)^2 + k has vertex (h, k).
Worked example: Put y = −3x^2 + 7x + 2 into vertex form. Factor the first two terms: y = −3(x^2 − (7/3)x) + 2. Half of −7/3 is −7/6, so add and subtract (7/6)^2 = 49/36 inside: y = −3[(x − 7/6)^2 − 49/36] + 2 = −3(x − 7/6)^2 + 3·(49/36) + 2 = −3(x − 7/6)^2 + 49/12 + 2 = −3(x − 7/6)^2 + 73/12. Vertex: (7/6, 73/12). Speed-check: inside the square we have x − 7/6, whose double is −7/3, matching the inside x-coefficient; and we added a·h^2 = (−3)·(49/36) outside with the correct sign when distributing. All neat, no lost socks.
Great question-this is exactly where completing the square feels fiddly. The reliable rule is: factor the leading coefficient a out of the x^2 and x terms first (even if a is negative). That’s not just a shortcut; it’s what lets you use the template (x + p)^2 = x^2 + 2px + p^2, which assumes the x^2 coefficient inside is 1. Once you have a(x^2 + (b/a)x + …), take half of (b/a), i.e., b/(2a), square it, and add–subtract that inside the bracket: y = a[x^2 + (b/a)x + (b/2a)^2] + c − a(b/2a)^2 = a(x + b/2a)^2 + c − b^2/(4a). Fractions do show up when b is odd; I usually accept them early to stay accurate. If you’re solving an equation and want to delay fractions, multiply both sides by 4a first so (2ax + b)^2 appears-neater for solving-but for vertex form, embrace the halves. Balance-wise, either add and subtract inside the same side (remembering the outside a multiplies what you added) or add the corresponding a⋅(b/2a)^2 to both sides-both are equivalent if you’re consistent. Quick check of your binomial: in a(x − h)^2 + k, you should have −2ah = b and k = c − a h^2 with h = −b/(2a). Clean vertex read-off: y = a(x − h)^2 + k means vertex (h, k) = (−b/(2a), c − b^2/(4a)). Example: y = −3x^2 + 5x − 7 = −3[x^2 − (5/3)x + (25/36)] − 7 + 25/12 = −3(x − 5/6)^2 − 59/12, so the vertex is (5/6, −59/12). Quick sanity check: −2a·h = −2(−3)(5/6) = 5, which matches b.