Completing the square is the bridge between ACT-level quadratic manipulation and AP-level antiderivative construction. Most candidates who reach an AP Calculus unit on u-substitution or the reverse chain rule still panic when the integrand is something like ∫ 1/(x²+4x+9) dx or ∫ √(−x²+6x−5) dx. The algebra looks hostile, the discriminant is negative, and the obvious move (factoring) simply does not work. In practice, the move that always works — the one that turns the ugly polynomial into a clean (x−h)²+k form — is the same completing-the-square routine taught in Algebra 2 and tested, in stripped-down form, on the ACT math section. The skill is not exotic. It is the same left-to-right rewrite a 14-year-old does on a quadratic homework sheet, applied to a 17- or 18-year-old's integration problem.
This article treats completing the square as a working tool, not as a vocabulary word. The goal is for an ACT-bound student — someone who is taking the ACT in the same academic year they sit AP Calculus, or who is auditing AP material to lift an ACT math score from 28 to 34 — to walk away with a four-step method, three integral families where the rewrite is non-negotiable, and a short list of the algebraic traps that cost the most points. The integration examples are real AP-style problems. The pacing commentary is anchored to the ACT's 60-minute, 60-question math section. By the end, completing the square should feel like a reflex, not a decision.
The four-step method: rewrite first, integrate second
Every completing-the-square flow in an AP integration problem reduces to four mechanical steps. The discipline is doing them in order, and writing the intermediate form on the page instead of doing the algebra in your head. Most integration errors at this level are not calculus errors. They are sign errors inside the square. I have seen students with full AP credit lose a six-point problem because they wrote (x+3)² when the correct half was (x−3)². Slow down, and the points stay on the page.
Step 1 — Isolate the quadratic and the linear terms. Pull the x² and x terms together, leaving constants and any other function of x to the side. For ∫ (x²+4x+5) dx, the work starts before the integral sign: rewrite the integrand as (x²+4x) + 5. For ∫ 1/(x²+6x+18) dx, you mentally split the constant 18 into “something to complete the square” and “something to leave over.” The split is what trips people up. Get this step right and the rest is bookkeeping.
Step 2 — Take half the linear coefficient and square it. If the linear term is bx, compute (b/2)². Add it inside a new perfect square and subtract it outside, so the value of the expression does not change. For x²+4x, the linear coefficient is 4, half of it is 2, and 2² is 4. The rewrite is x²+4x = (x+2)²−4. For x²+6x, you get (x+3)²−9. This step is identical to what the ACT math section tests when it asks you to rewrite a quadratic in vertex form — a question type that has appeared in some form on essentially every released ACT for a decade.
Step 3 — Fold the leftover constant back into the k of (x−h)²+k. In ∫ (x²+4x+5) dx, the leftover constant from step 2 is 5, so the full rewrite is (x+2)²+1. In ∫ 1/(x²+6x+18) dx, you folded −9 out, so you re-add it: x²+6x+18 = (x+3)²+9. This is the moment where sign slippage is most expensive. A single missed minus sign moves a problem from a clean arctan into an impossible integral. If you write the k as +9 in the denominator and it should be +9, fine. If you write +9 and it should have been something else, the antiderivative is wrong by a constant and you will not recover.
Step 4 — Integrate the rewritten form using a known family. With the square complete, the integral belongs to one of three families: a polynomial family, an arctan family, or an inverse-sine/arcsine family. The choice of family is dictated by the sign of the coefficient of the squared term and by whether the integrand is 1 over a square plus a constant, the square root of a constant minus a square, or a plain polynomial. The next three sections walk through each family in turn, with the worked steps and the ACT pacing note attached.
For ACT-bound students, the bridge to AP is even more direct than it looks. ACT math item 60-style problems — the last few questions in the math section, where the difficulty ceiling is highest — routinely test a vertex-form rewrite. The same move that earns a single ACT point on a quadratic rewrite earns a six- or seven-point AP sub-problem on an antiderivative. The economy is real: invest 90 seconds in completing the square once and the calculus writes itself.
Family 1: polynomial integrals after the rewrite
The polynomial family is the easy one, and it is where ACT students should build the habit. Consider ∫ (x²+4x+7) dx. Step 1 gives x²+4x. Step 2 gives (x+2)²−4. Step 3 folds the 7 back: (x+2)²+3. The integral becomes ∫ ((x+2)²+3) dx, which expands to ∫ (x+2)² dx + ∫ 3 dx. The first antiderivative uses the power rule with a u-substitution in your head: u = x+2, du = dx, so ∫ u² du = u³/3 = (x+2)³/3. The second antiderivative is 3x. Final answer: (x+2)³/3 + 3x + C. Two minutes of work, full credit.
The polynomial family is the right training ground because the only thing that can go wrong is the square. The integration itself is a power rule that an ACT student can do in their sleep. That is exactly the point: do the hard algebra carefully, and the calculus becomes trivial. For most candidates, this is the more efficient path than trying to integrate x²+4x+7 directly with the power rule on every term. Direct integration gives (x³)/3 + 2x² + 7x + C, which is correct but misses the training value. The completing-the-square version forces the (x+2) shift to stay visible, and that habit pays off the first time the problem has a square root or a 1/x in it.
A common variant is ∫ (2x²−12x+20) dx. Pull out the 2: 2 ∫ (x²−6x+10) dx. Inside the integral, complete the square on x²−6x: (x−3)²−9. Fold 10 back: (x−3)²+1. The integral becomes 2 ∫ ((x−3)²+1) dx = 2((x−3)³/3 + x) + C. Note the sign: half of −6 is −3, and (−3)² is 9, so the leftover is −9. A student who reflexively writes +9 will be off by 18 in the constant term, and the antiderivative will fail any derivative-check a grader applies. This is the exact error that the ACT math section catches in its medium-difficulty items: sign inside the square.
For pacing, treat every polynomial-family integral as 90 to 120 seconds on the AP. On the ACT side, the analogous vertex-form rewrite is a 30- to 45-second item at the back of the math section. Build the habit in ACT drills and the AP version is automatic. A practical drill: take any quadratic of the form x²+bx+c and rewrite it in (x−h)²+k form, then take its derivative and confirm you get 2(x−h). Ten repetitions a week for three weeks will retire the sign error for good.
Family 2: the arctan family — 1 over a sum of square and constant
The arctan family is the most-tested completing-the-square application in AP Calculus. The standard form is ∫ 1/(u²+a²) du, with antiderivative (1/a) arctan(u/a) + C. When the denominator is a quadratic with a positive leading coefficient and a negative discriminant, completing the square converts the denominator into exactly that shape. Take ∫ 1/(x²+4x+13) dx. Step 1: x²+4x. Step 2: (x+2)²−4. Step 3: x²+4x+13 = (x+2)²+9, with k = 9. Step 4: the integral is ∫ 1/((x+2)²+9) dx, and the form matches the template with u = x+2 and a² = 9, so a = 3. The antiderivative is (1/3) arctan((x+2)/3) + C.
The reason this matters for ACT students is that the discriminant check is the gate. A negative discriminant on the original quadratic tells you, before you even start the square, that the arctan family is in play. A zero discriminant tells you the denominator is a perfect square, and a partial-fraction or simple u-substitution is enough. A positive discriminant tells you to factor instead and use partial fractions. Reading the discriminant before integrating saves minutes per problem and is the single biggest pacing win at the back of the ACT math section, where items routinely present quadratics with no hint of which technique to use.
A second worked example: ∫ 1/(2x²+8x+20) dx. Factor the 2 out front: (1/2) ∫ 1/(x²+4x+10) dx. Complete the square on the inside: x²+4x+10 = (x+2)²+6. The integral is (1/2) ∫ 1/((x+2)²+6) dx. With a² = 6, a = √6, the antiderivative is (1/2)·(1/√6) arctan((x+2)/√6) + C = (1/(2√6)) arctan((x+2)/√6) + C. Notice that the (1/2) sitting outside does not disappear; it survives into the final coefficient. A student who folds it into the square — that is, who writes 2(x²+4x+10) = 2(x+2)²+2 and then tries to apply the template — will mis-apply the a. The correct move is to pull the leading coefficient outside, complete the square on what is left, and apply the template to a denominator of the form u²+a² only.
For the ACT math section, the same discriminant logic shows up as a 45-second item asking “which value of k makes the expression factorable” or “what is the minimum value of the quadratic.” Both are completing-the-square applications. Train the discriminant read in ACT drills and the AP arctan family is the same muscle at a higher speed.
Family 3: the arcsine family — square root of constant minus a square
The arcsine family is the mirror image of the arctan family, and it appears whenever the integrand is the square root of a constant minus a shifted square. The template is ∫ 1/√(a²−u²) du = arcsin(u/a) + C. On the AP exam this family tends to show up as a definite integral in a volume-of-revolution problem, or as an antiderivative step in a trig-substitution chain. The completing-the-square move is identical: pull the negative sign, complete the square on what is left, and check the form.
Worked example: ∫ √(−x²+6x−5) dx. Rewrite the radicand as −(x²−6x+5) = −((x−3)²−9+5) = −((x−3)²−4) = 4−(x−3)². So the integral is ∫ √(4−(x−3)²) dx. This is the standard a²−u² form with a² = 4 (so a = 2) and u = x−3. The trig-substitution that follows gives the antiderivative (x−3)/2·√(4−(x−3)²) + 2 arcsin((x−3)/2) + C. The mechanics of the substitution are not the point here. The point is that the radicand had to be rewritten as 4−(x−3)² before any of the trig machinery makes sense. Without the square, you are staring at −x²+6x−5 and assuming the problem is unsolvable.
For ACT students, the training value is reading the leading sign. A positive coefficient of x² with a negative constant term gives a downward parabola; the radicand is positive only between the roots, and a square-root integrand is well-defined there. A negative coefficient of x² with a positive constant term gives an upward parabola; the radicand is positive outside the roots, and an arcsine antiderivative (or its cousin, a hyperbolic substitution) is in play. The sign of the leading coefficient is the first thing to check, before completing the square, before deciding the family.
Common pitfall: students sometimes complete the square and then forget to flip the sign on the leftover constant. In the example above, x²−6x+5 = (x−3)²−4, so −(x²−6x+5) = −(x−3)²+4 = 4−(x−3)². The student who writes 4+(x−3)² has inverted the integrand and is now looking at the arctan family, not the arcsine family. The shape of the final antiderivative — arctan with a positive square, arcsin with a negative square — is the fastest cross-check.
ACT math overlap: where the skill is graded twice
The ACT math section tests completing the square in two recurring shapes, and AP Calculus reuses both shapes with an antiderivative stapled on top. The first shape is the vertex-form rewrite: “rewrite x²+6x+11 in the form (x+h)²+k.” The answer is (x+3)²+2. This is item 35-to-50 difficulty on a typical ACT, and it is a 30-second problem for a student who has practiced the half-the-coefficient-square-it routine. The second shape is the minimum or maximum value question: “what is the minimum value of f(x) = x²+8x+19?” The minimum is the k in the vertex form, which is 3, since x²+8x+19 = (x+4)²+3. The vertex itself is not asked, but the k is, and k falls out of the same completing-the-square flow.