On the AP Calculus AB and BC exams, a critical point is the single most common object a student is asked to locate, classify, and justify in a free-response question. It is the place where a continuous function's derivative is zero or undefined, and from it the whole behaviour of the function — increasing, decreasing, local extrema — falls out. Get the definition loose and you lose the line of reasoning that earns the point; get it tight and almost every optimisation, curve-sketching, and particle-motion prompt becomes easier to read. This article is the senior-tutor walkthrough of what the rubric actually demands when an AP Calculus prompt asks for a critical point, and how to write the justification in language that the readers recognise as complete.
The exam's working definition of a critical point
Most textbooks define a critical point of f as a number c in the domain of f such that f '(c) = 0 or f '(c) is undefined. The AP Calculus Course and Exam Description tightens this further: a critical point only counts if the function is actually defined at that x-value. A candidate point where the derivative fails to exist because the function itself is not defined there is not, for AP scoring purposes, a critical point — it is a point that is not in the domain and therefore cannot host a local extremum.
The most common student mistake is to write down the zeros of the derivative and call them all critical points without checking the domain. A polynomial's critical points are exactly its real zeros of f ', but a rational function, a piecewise function, or any function with restricted domain will throw away a point the moment the original function is undefined there. On the exam, this is exactly the distinction the readers are looking for in the first rubric line.
The BC exam adds a second wrinkle: critical points can be found for parametric, vector, and polar functions as well as the usual y = f(x) form. For a parametric curve (x(t), y(t)), a critical point is a value of t where dy/dx = 0 and dx/dt ≠ 0, or where dx/dt = 0 and dy/dt ≠ 0. Forgetting to add the “and the other derivative is not zero” condition is a classic place to give back a point, because vertical tangent behaviour gets folded into the same line of reasoning and the rubric treats them differently.
Worked micro-example. Let f(x) = (x − 1)² / x on its natural domain. A student who jumps straight to f '(x) = 0 finds x = 1 as the only candidate, but the function is undefined at x = 0, so the rubric expects a critical point at x = 0 as well — even though the derivative is not zero there. A complete answer: critical points at x = 0 and x = 1, with a one-line note that f is undefined at x = 0.
This domain-first habit is the single biggest separator between a 4 and a 5 on classification prompts. Hold on to it as you read the rest of the article.
Three question shapes that ask for a critical point
Free-response prompts that involve critical points fall into a small number of recognisable shapes. Identifying the shape before you start writing is half of the work, because each shape has its own rubric line and its own order of operations.
Shape one: the open-interval extractor. The prompt gives f on an interval (a, b) and asks for the location and value of all absolute extrema. The solution must (i) find the critical points, (ii) evaluate f at each, (iii) compare with the function's behaviour at the endpoints, and (iv) state the maximum and minimum. Most students lose the point not on the calculus but on the order: they jump to the conclusion before the rubric sees the comparison step.
Shape two: the local-behaviour classifier. The prompt asks for the intervals on which f is increasing or decreasing, or for the location and nature of relative extrema. Here the critical points are intermediate objects — the x-values that mark the boundaries of the sign chart. The first derivative test or the second derivative test does the classification, but the rubric still wants the critical point list written down explicitly. A common error is to skip the list and go straight to “f is increasing on (a, b)”, which loses the row of points that asks for the candidate values.
Shape three: the applied-optimisation prompt. A word problem — minimise cost, maximise area, minimise time — is reduced to a single-variable function and then handed to the same machinery. The critical point is the candidate answer, and the rubric demands a closing sentence confirming that it is a maximum or minimum, often with a one-sentence reason (concavity, endpoint comparison, or second derivative test). Forgetting the confirmation sentence is the most common reason an otherwise correct optimisation answer scores 5 out of 6 instead of 6 out of 6.
Multiple-choice questions compress these shapes further. A typical stem gives a graph and asks for the x-value of a local maximum; the answer is the critical point where the tangent is horizontal and the function changes from increasing to decreasing. A more pointed MCQ might list four candidate x-values, only one of which is a critical point by the rubric's definition — for instance, a corner where the function is defined but the derivative is undefined, paired with a stationary inflection that is not a local extremum. The same definition; a quicker read.
Finding critical points: a procedure that earns the rubric line
The reason most students lose the “find the critical points” line is not that they cannot find them — it is that the reader cannot tell, from the work on the page, which values are being claimed and why. The AP readers reward a specific kind of writing: claim, compute, justify, in that order, with each step visible. The procedure below is the one I teach in the AP Calculus BC review block the week before the exam, and it transfers cleanly to AB.
Step 1. State the domain. One short sentence: “f is defined for all real x except x = −1” or “the domain of f is [0, 8].” This anchors every later step. The reader can now see whether a candidate is in or out of bounds, and the student has a check against the trap of including a non-domain zero of the derivative.
Step 2. Differentiate and simplify. Show the derivative in factored or quotient form. The rubric does not require a particular algebraic form, but the cleaner the form, the easier the sign chart in step 4. Avoid leaving a derivative as a single unfactored polynomial unless that is the only practical form.
Step 3. Set f '(x) = 0 and solve. Write down each root. This is the list of candidate critical points where the derivative is zero. If the prompt is on the BC exam, also state the values of x where f ' is undefined (vertical tangents, absolute-value corners, denominators zero while the original function is defined).
Step 4. Cross out anything outside the domain. If the prompt gives a closed interval, this is where the endpoints enter — they are candidates for absolute extrema, but they are not critical points. If the prompt is on the whole real line, then any candidate where f itself is undefined is removed from the list.
Step 5. State the critical points in plain language. “The critical points of f are x = a, x = b, and x = c.” This is the line the rubric wants to see. It is short, complete, and easy for the reader to award points against.
The procedure is intentionally mechanical. In my experience, students who skip steps — usually the domain statement and the explicit list — give back a point per FRQ on a topic that is otherwise worth full marks. The five lines cost a minute. The point costs a score band.
Classifying critical points: first derivative test versus second derivative test
Once the critical points are written down, the next rubric line asks the student to decide which ones are local maxima, which are local minima, and which are neither. Two tests are taught; both are accepted by the rubric. Choosing between them is partly preference and partly signal: the second derivative test is faster but only works when the second derivative is easy to evaluate and is non-zero at the critical point. The first derivative test is slower but never fails, as long as the sign chart is written out.
First derivative test. Build a sign chart of f '(x) on either side of each critical point. If f ' changes from positive to negative, the point is a local maximum. From negative to positive, a local minimum. No sign change, no local extremum — the point is stationary but not extremal, which on the AP exam is often phrased as “a critical point that is not a local extremum.”
The rubric wants the chart drawn or the sign-change written down, not just the conclusion. A line that says “f has a local max at x = 2 because f ' changes from positive to negative” is enough. A bare “local max at x = 2” leaves the reader no evidence to award the justification point.
Second derivative test. Evaluate f '' at each critical point. If f ''(c) > 0, the function is concave up at c and the critical point is a local minimum. If f ''(c) < 0, concave down and a local maximum. If f ''(c) = 0 or the test is otherwise inconclusive, the rubric allows — and in fact expects — a fallback to the first derivative test for that point. Students who write “the second derivative test is inconclusive” and stop there lose the row; the rubric reads that as a non-answer.
Worked micro-example. f(x) = x⁴ − 4x³. f '(x) = 4x³ − 12x² = 4x²(x − 3). Critical points at x = 0 and x = 3. At x = 3, the second derivative is f ''(3) = 36 > 0, so x = 3 is a local minimum. At x = 0, the second derivative is 0, so the test is inconclusive; the sign chart of f ' shows that the derivative is negative on either side of 0, so x = 0 is a critical point that is not a local extremum. The complete FRQ sentence: “f has a local minimum at x = 3 with value f(3) = −27, and a critical point at x = 0 that is not a local extremum.”
Notice that the answer includes the y-value at the local extremum, which the rubric often asks for in the same row as the classification. Pairing the two is a small habit that catches one extra point on almost every classification prompt.
Critical points versus inflection points: how the rubric keeps them separate
An inflection point is not a critical point, and the AP exam will test the distinction whenever a graph or a derivative graph is on the page. An inflection point is a place where the concavity of the function changes; a critical point is a place where the derivative is zero or undefined. The two overlap only at a stationary inflection — a point that is both critical and inflectional, like x = 0 in the previous example.
Where students lose the point is at the level of vocabulary. The rubric awards the row for “local minimum” and the row for “inflection point” separately, and a student who calls a local minimum an inflection point loses the first row even if the x-value is correct. The cure is a one-sentence mental check before answering: at this point, is the derivative zero or undefined? If yes, it could be a critical point. Is the concavity changing? If yes, it is also an inflection point. The two questions are independent.
The multiple-choice section exploits this. A typical prompt: “At x = c, f '(c) = 0 and f '' changes sign. Which of the following must be true?” The correct answer is that (c, f(c)) is a point of inflection; the answer “f has a local extremum at c” is a trap, because the second derivative changing sign does not by itself force an extremum. The same logic, in reverse, is on a sign-chart prompt where the derivative is zero and the function changes concavity but not monotonicity.
Students preparing for the AP Calculus exam should treat the critical point / inflection point distinction as a vocabulary drill, not a calculus drill. The calculus is the same; the labels are what the rubric scores.
The FRQ rubric: how a critical-point line of reasoning is scored
The readers for AP Calculus work from a rubric written before the exam, and the rubric for a critical-point prompt is consistent across years. There are typically three rows: (i) finding the critical points, (ii) classifying them with a stated reason, and (iii) stating the conclusions in context (for applied problems) or as a list of intervals (for curve-sketch problems). Each row is one or two points. Skipping a row — or performing the row's work without writing it down — is how students leave points on the table.
The first row awards credit for an explicit list of critical points, including any where the derivative is undefined. The rubric's wording is something like: “The student finds all critical points of f in the given interval.” A student who finds three of four critical points earns partial credit on that row, usually 1 out of 2, and that partial credit is what separates a 4 from a 5 on a single FRQ.
The second row awards credit for the classification with a reason. “Local maximum at x = a because f ' changes from positive to negative” is a full-credit sentence. “Local maximum at x = a” alone is half credit at best; “critical point at x = a” with no classification is no credit on that row.
The third row is the one students forget. For an applied optimisation problem, the rubric wants a closing sentence such as: “The minimum cost occurs when x = 5, giving a minimum cost of $320.” For a curve-sketch problem, the rubric wants the intervals of increase and decrease written out, with the critical points as endpoints. A correct classification that is never connected back to the question loses the third row and roughly one point of FRQ credit.
The total for a critical-point FRQ is usually 6 to 9 points, depending on whether the problem also asks for concavity or absolute extrema. A student who nails the procedure above will score full credit on the critical-point portion; the rest of the question is then just additional rows on the same page.
| Rubric row | What the reader is looking for | One-sentence template |
|---|---|---|
| Find the critical points | Explicit list, domain respected | “The critical points of f on [0, 8] are x = 1, 3, 5.” |
| Classify with a reason | Sign change or second derivative sign | “x = 3 is a local minimum because f ' changes from negative to positive.” |
| State the conclusion in context | Answer the question that was asked | “The minimum value of f on [0, 8] is f(3) = −2.” |
| Justify domain choices | Endpoints, undefined values, restricted interval | “x = −1 is not a critical point because f is undefined there.” |
Common pitfalls and how to avoid them
Across several years of marking practice sets, the same handful of mistakes account for most of the lost points on critical-point prompts. None of them are hard to fix; they are habits, and habits can be replaced in the two weeks before the exam.
- Confusing “critical point” with “zero of the derivative”. The fix is the domain step. Spend ten seconds on it; the rubric's first row is built around the result.
- Calling an inflection point a critical point. Vocabulary, not calculus. Drill the two definitions side by side until they sit in different boxes.
- Skipping the sign-change justification. The reader cannot infer the reason from the conclusion. Write one short clause after each classification.
- Forgetting the closing sentence on applied problems. The rubric's last row is often the most generous. A single sentence recovers the point.
- Including endpoints in the critical-point list on a closed-interval problem. Endpoints are candidates for absolute extrema, not critical points. Keep the two lists separate on the page.
- Using the second derivative test when f ''(c) = 0 and stopping. The rubric expects a fallback. Practise the phrase “the second derivative test is inconclusive, so we use the first derivative test” until it is automatic.
Worked example: a typical AP-style critical-point FRQ
The prompt: let f(x) = x² √(4 − x²). Find the critical points of f on its domain, classify each as a local maximum, local minimum, or neither, and state the absolute maximum and minimum of f on its domain.
Step 1. Domain. The square root requires 4 − x² ≥ 0, so the domain is [−2, 2].
Step 2. Derivative. Write f(x) = x² (4 − x²)¹ᐟ². By the product rule, f '(x) = 2x (4 − x²)¹ᐟ² + x² · (1/2)(4 − x²)⁻¹ᐟ² · (−2x). Simplify: f '(x) = 2x (4 − x²)¹ᐟ² − x³ / (4 − x²)¹ᐟ². Put over a common denominator: f '(x) = (2x(4 − x²) − x³) / (4 − x²)¹ᐟ² = (8x − 2x³ − x³) / (4 − x²)¹ᐟ² = x(8 − 3x²) / (4 − x²)¹ᐟ².
Step 3. Zeros of the numerator. x = 0 or 8 − 3x² = 0, giving x = ±√(8/3) = ±(2√6)/3. Both are inside the domain [−2, 2] because 2√6/3 ≈ 1.63 < 2.
Step 4. Undefined values. The denominator is zero when x = ±2. At those points, the original function is defined (f(±2) = 0), and the derivative is undefined. So x = ±2 are critical points. The function is also defined on the closed interval, so the endpoints are candidates for absolute extrema but are not, in the textbook sense, critical points. The rubric accepts either convention here; the safer choice on the AP exam is to list x = ±2 as critical points (derivative undefined) and add a note that they are also endpoints.
Step 5. List. The critical points are x = −2, −(2√6)/3, 0, (2√6)/3, 2.
Step 6. Classify. Use the first derivative test by sign. On the domain from left to right: f ' is positive just right of x = −2 (the numerator x(8 − 3x²) is positive for small negative x near −2), then it crosses zero at x = −(2√6)/3 and becomes negative, so that point is a local maximum. It crosses zero again at x = 0 and becomes positive, so x = 0 is a local minimum. It crosses zero at x = (2√6)/3 and becomes negative, so that point is a local maximum. At x = ±2, f = 0 and f is non-negative everywhere on the domain, so the values at the endpoints are absolute minima.
Step 7. Conclude. Local maxima at x = ±(2√6)/3 with value f((2√6)/3) = (8/3)√(4 − 8/3) = (8/3)·√(4/3) = (8/3)·(2/√3) = (16)/(3√3) = (16√3)/9. Local minimum at x = 0 with value 0. Absolute minimum on the domain is 0, attained at x = 0 and at the endpoints x = ±2. Absolute maximum on the domain is (16√3)/9, attained at x = ±(2√6)/3.
Total rubric credit: 9 points on a typical critical-point-and-extrema prompt. The student who writes the seven steps above in the order shown earns full credit. The student who skips the domain step, omits the sign-change reasons, or forgets the closing sentence loses 2 to 3 points on the same content.
Putting it together: a two-week preparation plan
Critical points are a high-frequency, high-payoff topic. A focused two-week block lifts most students from “I can find them” to “I can score the full row on the rubric.” The block below is what I run in the AP Calculus AB and BC review weeks at AP Courses.
Days 1 to 3. Definition drill. State the textbook definition of a critical point from memory, then state the AP-specific version that adds the domain condition. Do ten flashcard prompts: given a function, list the critical points. Include at least three rational functions, two with restricted domains and one with a removable hole.
Days 4 to 6. Classification drill. Take a sign chart of f ' from a textbook problem and write the local extrema with a one-clause reason. Switch to the second derivative test on a separate set of polynomial problems. Practise the inconclusive case: when f ''(c) = 0, write the fallback sentence and the first-derivative-table that resolves it.
Days 7 to 9. FRQ integration. Pull three released FRQs that include a critical-point line. Time yourself: 15 minutes per question. Score your own work against the rubric. The point of the exercise is not the calculus — it is the writing. Get used to producing the seven-step procedure in the form the rubric wants.
Days 10 to 12. Vocabulary check. Sort a mixed list of points into critical, inflection, both, or neither. Use graphs, derivative graphs, and tables. The exam's MCQ section rewards the student who can do this in under 30 seconds per item.
Days 13 to 14. Mock and review. Sit a full MCQ section with a stopwatch. Aim for under 90 seconds per critical-point item, including reading time. Then sit one full FRQ that contains a critical-point prompt and score it against the published rubric.
For most students reading this, the lift comes not from learning new calculus but from learning the rubric's vocabulary. A clean, complete answer reads as three or four short sentences, and a clean, complete answer is what the AP readers are trained to award points to.
Critical points sit at the centre of the AP Calculus AB and BC exam's free-response section, and the rubric around them is unusually stable. Practise the seven-step procedure, drill the vocabulary, and the row of points on classification prompts becomes a routine earn. AP Courses' AP Calculus BC programme runs each student through three released FRQs a week with rubric scoring, turning the critical-point row from a soft 5 into a guaranteed 6.