6 AP Calculus limit questions that separate a 4 from a 5
A senior tutor's walkthrough of AP Calculus infinite limits and limits at infinity: worked free-response examples, end behaviour models, and the errors that cost points.
In this article (12 sections)
Two phrases that look almost identical on the page behave like completely different creatures in AP Calculus: infinite limits and limits at infinity. The first asks what happens to a function as the input approaches a finite value and the output shoots off without bound. The second asks what happens to a function as the input itself grows without bound. Conflating them is one of the most common reasons a candidate loses points on the multiple-choice section of the AP Calculus AB or BC exam, and the misunderstanding tends to cascade into the free-response questions on asymptotic behaviour, end behaviour models, and improper integral convergence.
Defining infinite limits: when the function blows up near a finite point
An infinite limit is a statement about the behaviour of f(x) as x approaches a specific, finite value. The notation lim x→a f(x) = ∞ does not mean the limit exists; it is shorthand for the fact that f(x) can be made arbitrarily large by choosing x close enough to a. Two distinct one-sided behaviours are possible. The right-hand limit lim x→a+ f(x) might equal +∞ while the left-hand limit lim x→a- f(x) equals -∞, in which case the two-sided limit is said to not exist because the function heads off in opposite directions. The classic example is the reciprocal function 1/x near zero: from the positive side the values climb without bound, from the negative side they descend without bound. Students are often tempted to write lim x→0 1/x = ∞ as a single statement, which loses the one-sided information that AP graders specifically look for in free-response justifications.
Vertical asymptotes emerge directly from this family of limits. A vertical asymptote at x = a is a line that the graph approaches but never touches, and it exists whenever at least one of the one-sided limits is infinite. The test for locating candidate asymptotes is mechanical: factor the denominator, find its real zeros, and check each zero in the original (unreduced) expression. A zero that survives cancellation is a hole in the graph, not an asymptote. The function (x² - 1)/(x - 1), for instance, simplifies to x + 1 with a removable discontinuity at x = 1, while 1/(x - 1) keeps an honest asymptote at the same point. The most common error I see in private tutoring is a student who cancels first and then forgets to record the hole, costing a full point on a free-response sub-part. Write the zero, decide, and only then move on.
Worked example: locating infinite limits algebraically
Consider f(x) = (x + 2)/(x - 3)². The denominator vanishes at x = 3 and the numerator does not, so the immediate suspicion is an infinite limit. Direct substitution gives the form 5/0, which is not a number and not a valid answer. To classify the behaviour, factor the squared term: as x approaches 3 from either side, the squared denominator is positive and approaching 0, while the numerator approaches 5, a positive constant. The quotient of a positive number and a vanishingly small positive number grows without bound from both sides, so lim x→3 f(x) = +∞ and the two-sided limit is recorded as +∞ rather than as 'does not exist' in the ordinary sense. If the denominator had been raised to the first power instead of squared, the analysis would branch: approaching from the left, (x + 2) is positive and (x - 3) is negative, producing a negative quotient that descends to -∞, while approaching from the right the quotient climbs to +∞. The two one-sided infinities would then have opposite signs, and the two-sided limit would be declared not to exist. That distinction between '+∞' and 'DNE' is worth one full point in most AP free-response scoring rubrics.
Defining limits at infinity: tracing end behaviour as x grows
A limit at infinity is a different question. The notation lim x→∞ f(x) = L describes the horizontal destination of the graph as x becomes arbitrarily large, and the answer L is typically a finite real number, not infinity itself. The same machinery is used for x→ -∞, with attention paid to whether the function is even, odd, or neither, because that parity determines whether the two one-sided end behaviours agree. Polynomials, rational functions, root functions, exponential functions, and logarithmic functions all have well-defined end behaviours that can be read directly from their dominant term.
For polynomials, the limit at ±∞ is governed entirely by the leading term. If the leading term has odd degree with a positive coefficient, lim x→+∞ f(x) = +∞ and lim x→-∞ f(x) = -∞. If the degree is odd with a negative leading coefficient, both signs flip. If the degree is even, both one-sided limits are equal: they are +∞ if the leading coefficient is positive and -∞ if negative. The polynomial 3x⁴ - 2x² + 7, for example, has lim x→±∞ f(x) = +∞ because the leading term dominates for both signs of x. There is no horizontal asymptote in this case; the function climbs to +∞ on both sides.
Horizontal asymptotes correspond to finite limits at infinity. The function f(x) = (2x + 1)/(x - 5) has lim x→±∞ f(x) = 2 because the leading coefficients of numerator and denominator, 2 and 1, form a ratio of 2. The precise rule is that for a rational function with numerator degree equal to denominator degree, the horizontal asymptote is the ratio of leading coefficients. If the numerator degree is strictly less than the denominator degree, the horizontal asymptote is the x-axis itself (y = 0). If the numerator degree is strictly greater, there is no horizontal asymptote, and the function diverges to ±∞ in the direction of the leading term's sign. These three cases are pure pattern recognition once the degrees are compared, and pattern recognition is exactly the kind of fast classification that AP students need to perform inside the roughly two minutes available per multiple-choice question.
End behaviour models: the 'Big O' shortcut for free-response
AP free-response frequently asks for the end behaviour model of a function, phrased as 'identify a function g(x) such that f(x) ≈ g(x) as x → ±∞'. The rule is simple: divide numerator and denominator by the highest power of x appearing in the denominator, then drop every term that vanishes in the limit. The surviving expression is the end behaviour model. For f(x) = (3x³ - 5x)/(x² + 1), dividing top and bottom by x² yields (3x - 5/x)/(1 + 1/x²), and as x → ∞ the terms 5/x and 1/x² both head to 0, leaving 3x as the model. The end behaviour of 3x is +∞, so the end behaviour of f is +∞, and there is no horizontal asymptote. The model is useful beyond end behaviour: if a question asks for the slope of the asymptote, the model gives the answer directly. For most candidates I work with, practising three to four of these reductions in a single sitting is enough to make the pattern automatic, and once it is automatic the time saved on the free-response section is measurable in minutes.
For your AP prep
Side-by-side comparison: how the two limit families differ on the AP exam
Placing the two concepts next to each other is the fastest way to lock the distinction into long-term memory. The table below summarises the structural differences that the AP exam routinely probes.
| Feature | Infinite limit lim x→a f(x) | Limit at infinity lim x→∞ f(x) |
|---|---|---|
| What x is doing | Approaching a finite number a | Growing without bound |
| Typical answer | ±∞ or DNE | A finite real number L, or ±∞ |
| Geometric feature | Vertical asymptote at x = a | Horizontal asymptote y = L |
| Algebraic test | Substitute a; look for /0 with non-zero numerator | Compare degrees of numerator and denominator |
| One-sided analysis | Required: signs may differ | Required only when end behaviours differ in sign |
| Common error | Writing '= ∞' for the two-sided limit | Forgetting the leading-coefficient ratio when degrees match |
The single highest-leverage habit a student can build is to read the arrow. If the arrow points to a number, the question is about an infinite limit and the answer is about how the function behaves near that number. If the arrow points to ∞, the question is about end behaviour and the answer is about how the function behaves far from the origin. AP items are written deliberately to make the arrow the deciding clue, and candidates who pause for half a second to read the arrow rarely mix up the two cases.
Worked free-response style problems: from infinite limits to limits at infinity
The AP Calculus free-response section often contains a sub-part that moves from one limit family to the other in the same problem, and this is where the separation between 4-scoring and 5-scoring candidates tends to widen. A typical stem might give f(x) = (x² - 4)/(x - 2) and ask for three sub-parts: the limit as x approaches 2, the limit as x approaches infinity, and the location of any asymptotes. Working through it step by step, the first limit is the indeterminate form 0/0, so the function must be simplified before evaluation. Factoring the numerator as (x - 2)(x + 2) and cancelling the common factor with the denominator leaves x + 2, and substituting x = 2 gives 4. There is no infinite limit here, despite the apparent denominator zero; the zero was a removable discontinuity, and the candidate who records an asymptote at x = 2 has lost a point.
The second sub-part, lim x→∞ (x² - 4)/(x - 2), behaves differently. The numerator has degree 2 and the denominator degree 1, so there is no horizontal asymptote: the simplified form x + 2 is itself the end behaviour model, and the limit is +∞. The third sub-part, on asymptotes, should record that there is no horizontal asymptote and explicitly note the absence of a vertical asymptote at x = 2. The complete answer is a tight three-sentence paragraph that names the removable discontinuity, states the end behaviour, and uses the phrase 'no vertical asymptote' rather than leaving the topic silent. AP graders award points for the explicit negative claim, not for the implicit absence.
A second worked example uses g(x) = (sin x)/x, which is a classic limit at infinity problem. The numerator is bounded between -1 and 1 for all real x, while the denominator grows without bound. By the squeeze theorem, lim x→±∞ (sin x)/x = 0. The function has no vertical asymptote: it is undefined only at x = 0, and there lim x→0 (sin x)/x = 1, so that discontinuity is removable. The horizontal asymptote y = 0 appears on both sides, and the end behaviour model is 0. Candidates who reflexively try to apply polynomial degree rules here will mis-classify the function, because the squeeze theorem is the correct tool for non-rational functions. The lesson is that the algebraic shortcuts for rational functions are powerful but bounded: when the function is not rational, return to first principles.
Common pitfalls and how to avoid them
In the practice sets I review, the same handful of errors account for the majority of lost points. Listing them explicitly is more useful than describing them generically, because each one has a specific fix.
- Conflating the two limit families: writing lim x→∞ f(x) = ∞ when the question asked for an infinite limit near a finite point, or vice versa. The fix is to read the arrow first and announce the family aloud before computing.
- Ignoring one-sided behaviour: a two-sided infinite limit exists in the +∞ sense only if both one-sided limits agree in sign. The fix is to always check the sign of the numerator and the sign of the factor that is approaching zero separately.
- Cancelling before checking for holes: removing a common factor and then forgetting to record the removable discontinuity. The fix is to write the cancellation step and append a one-line note: 'hole at x = a, no asymptote'.
- Misapplying the degree rule: assuming a horizontal asymptote exists whenever a function is rational, even when the numerator degree is higher. The fix is to compare degrees first; no comparison, no conclusion.
- Confusing the end behaviour model with the function itself: writing the model as if it were the original f(x) on a free-response. The fix is to use the '≈' symbol explicitly and to state the direction (x → ∞ or x → -∞) in which the approximation holds.
Tactical study plan for the next four weeks
A focused four-week study plan works best when it builds the underlying skill, rapid classification under time pressure, before adding volume. Weeks 1 and 2 should be dedicated to limit family recognition: ten minutes per day on locating vertical asymptotes, ten minutes on horizontal asymptotes, and ten minutes on end behaviour models. Weeks 3 and 4 should integrate free-response style multi-part problems, timed at the exam's pace of about fifteen minutes per free-response question, so that classification becomes automatic.
Connecting asymptote reasoning to exam format awareness
Candidates who understand the scoring weights of the multiple-choice and free-response sections, and the sub-weights within the free-response, tend to allocate their time during the exam more effectively. The AP Calculus AB and BC exams last 3 hours 10 minutes: Section I has 42 multiple-choice questions in 1 hour 40 minutes (Part A, 29 questions in 62 minutes without a calculator; Part B, 13 questions in 38 minutes with a graphing calculator), and Section II has 6 free-response questions in 1 hour 30 minutes, each scored out of 9 points. Each section is worth 50% of the exam score, so the free-response section is weighted slightly more heavily per minute. The implication is that a candidate should not linger on a multiple-choice item far past its two-to-three-minute share, even if it means flagging it for return, because the free-response minutes are the higher-leverage minutes.
Reading the prompt
In AP Calculus, reading the limit notation correctly is the difference between a correct answer and a wrong one. The exam rewards the candidate who takes an extra beat to classify the task before producing output, and penalises the candidate who rushes into a response that addresses the wrong question. The highest-leverage change to make this week is to add a five-second pause before every practice response and to use that pause to name the limit family out loud.
Conclusion and next steps
Infinite limits and limits at infinity are distinct families of AP Calculus questions that share a name, a symbol, and a propensity for confusion. Mastering the distinction requires reading the arrow, classifying the task, and applying the right algebraic shortcut for rational functions or the right theorem (such as the squeeze theorem) for non-rational ones. A diagnostic AP Calculus quiz focused on asymptote and end behaviour items is the natural starting point for candidates who want to convert classification skill into a higher score.