The ratio test is one of the workhorse tools in the AP Calculus BC series and convergence unit, and it shows up on the exam in a small number of predictable shapes. Most students remember the headline rule — compute the limit of the absolute value of aₙ₊₁ over aₙ, then read off convergence or divergence from that limit — but the rubric actually scores the work in three distinct rows: the limit itself, the conclusion that follows from it, and the justified handling of the L = 1 case when the limit lands there. Each of those rows is awarded independently, and the test writers know exactly which row students fumble.
For AP Calculus BC candidates preparing for the FRQ section, the ratio test is rarely the entire question. It typically appears as a sub-step inside a longer problem about a power series, a Taylor polynomial, or an alternating series that must be evaluated alongside other convergence tests. The exam format gives roughly six FRQs, each worth nine points, and convergence/series work appears in at least one of them on most administrations. The MCQ section also rewards the same fluency at speed, since a poorly chosen ratio-test computation can cost 90 seconds you do not have. Scoring at the 5 level on this content means treating the ratio test as a structured argument with named rows, not as a single formula applied by feel.
What follows is a preparation-focused walkthrough: when the ratio test is the right tool, what each line of the rubric actually scores, and how to handle the three trouble cases — L < 1, L > 1, and the inconclusive L = 1 — without losing points on either section of the AP exam.
What the ratio test actually says, in the form AP Calculus scores
The ratio test is a statement about an infinite series ∑aₙ built from three explicit cases. Compute L = lim |aₙ₊₁ / aₙ| as n approaches infinity. If L < 1, the series converges absolutely. If L > 1, the series diverges — and in this case you can also conclude that aₙ does not tend to zero, which is information the rubric often rewards in a follow-up row. If L = 1, the test is inconclusive and you must switch to a different convergence argument: comparison, limit comparison, the integral test, or the alternating series test, depending on the sign pattern of aₙ.
The AP Calculus BC syllabus lists the ratio test explicitly as a required tool, and the FRQ scoring guidelines treat it as a multi-line argument. Three rows are visible in published rubrics for ratio-test problems:
- The setup row: the candidate has written |aₙ₊₁ / aₙ| with the correct substitution of n+1 for n in the original expression for aₙ, and has begun to simplify.
- The limit row: the limit has been evaluated correctly, typically by cancelling common factors, applying properties of exponents, or recognising a limit of the form lim (n/(n+1))ᵏ = 1.
- The conclusion row: the candidate has stated the correct verdict — converges absolutely, diverges, or inconclusive — and tied it explicitly to the value of L.
Notice what is not on that list: the value of aₙ itself, the radius of convergence of a related power series, or the sum of the series. The ratio test on the AP exam almost never asks for the sum; it asks for a verdict. A candidate who correctly computes L = 1/2 but writes 'the series converges to some value S' has earned only the limit row, not the conclusion row. The two rows score independently, and the conclusion must reference L explicitly to score the second point.
For most candidates, the L < 1 case is the easiest to write, and the L = 1 case is the easiest to mis-write. In my experience marking practice FRQs, the single most common error is declaring the series divergent when L = 1 simply because the candidate expected a clean answer. The ratio test never licenses that conclusion, and the rubric charges a point for it.
When to choose the ratio test on an AP Calculus FRQ
The ratio test is the right choice when aₙ contains factorials, exponentials in n, or products of consecutive integers. These are the structural signatures the test writers reach for, and recognising them is roughly half of method selection. A series like ∑ n! / 5ⁿ, ∑ (3ⁿ) / n!, or ∑ 1 / (2n choose n) is practically begging for the ratio test — the algebra collapses cleanly, and L lands at a number far from 1. By contrast, a series like ∑ 1/n² or ∑ sin(1/n) is a poor fit; the ratio test produces L = 1 and forces a method switch, wasting minutes that should have been spent on a p-series or comparison argument in the first place.
The MCQ section tests this selection skill under time pressure. On the 45-question multiple choice part of the exam, ratio-test items typically pair with one or two comparison-test items, and the candidate who runs the ratio test on a p-series loses 60 to 90 seconds per item — fatal at the back end of the section. The triage rule I teach is straightforward: if the dominant operation in aₙ is multiplication across n (factorials, powers, binomial coefficients), reach for the ratio test; if the dominant operation is addition or a function of n (sin, cos, polynomials in n), start with nth term, comparison, or integral test.
Structural signatures that point to the ratio test
Three families appear repeatedly on the exam. The first is aₙ = (polynomial in n) · rⁿ, where r is a constant. The ratio collapses the polynomial ratio to 1 and leaves L = |r|. The second is aₙ involving n! in either the numerator or denominator. Factorials dominate exponentials with overwhelming force — n! grows faster than cⁿ for any fixed c — so L lands at 0 or infinity, both of which produce decisive verdicts. The third is aₙ built from binomial coefficients such as (2n choose n) or (n + k choose k); the ratio simplifies to a rational function of n that often has a clean limit.
Two families are poor matches even though students often try them. The first is geometric-like series where aₙ = rⁿ; the ratio test works (L = |r|), but the geometric series test is faster and scores the same point. The second is alternating series; the ratio test still applies to |aₙ|, but the alternating series test is almost always simpler and gives additional information about error bounds that the exam sometimes asks about. A good rule of thumb: if the alternating series test is available, use it; reserve the ratio test for terms where the alternating structure is not the dominant feature.
The L < 1 case: writing a clean convergence argument
When L < 1, the rubric awards the conclusion point for a single explicit sentence: 'Since L = [value] < 1, the series converges absolutely by the ratio test.' Two ingredients must appear. The value of L must be stated as a number, not as an expression left in limit form. The comparison L < 1 must be written out — simply asserting 'converges' without the inequality is a common error that costs the second row.
A worked example. Let aₙ = n² / 3ⁿ. Compute |aₙ₊₁ / aₙ| = ((n+1)² / 3ⁿ⁺¹) · (3ⁿ / n²) = (n+1)² / (3 n²). Take the limit: as n → ∞, (n+1)² / n² → 1, so L = 1/3. Write: 'L = 1/3 < 1, so the series converges absolutely by the ratio test.' That is the entire argument; it scores the limit row and the conclusion row on a typical AP rubric.
Three tactical notes on this case. First, do not bother computing the actual sum of the series; the ratio test never produces it, and time spent hunting for a closed form is time stolen from other problems. Second, the absolute value bars in the limit definition matter for alternating series but are visually redundant for positive-term series; either way, include them on the exam for safety. Third, when the limit involves a ratio of polynomials or exponentials, cancel common factors inside the limit expression before taking n → ∞ — the test writers award partial credit on the limit row for visible simplification, not only for the final number.
The L > 1 case: divergence, and the bonus row
When L > 1, the rubric awards two conclusion points rather than one, because the ratio test in this case gives a second piece of information: the terms aₙ do not tend to zero. Many AP FRQs on the divergence case include a follow-up row that asks whether the series 'meets the necessary condition for convergence' or 'fails the nth term test', and the ratio test supplies that answer immediately.
Worked example. Let aₙ = nⁿ / n!. Compute |aₙ₊₁ / aₙ| = ((n+1)ⁿ⁺¹ / (n+1)!) · (n! / nⁿ) = (n+1)ⁿ / nⁿ · (n+1)/(n+1) = (1 + 1/n)ⁿ. The limit as n → ∞ of (1 + 1/n)ⁿ is e ≈ 2.718. Write: 'L = e > 1, so the series diverges by the ratio test; in particular, aₙ does not approach 0.' The phrase 'does not approach 0' is what scores the bonus row on most published rubrics.
This is a useful pattern to memorise. Whenever the ratio test gives L > 1, the candidate is licensed to make a stronger claim than simple divergence — namely, that the terms fail the divergence test. If a follow-up FRQ part asks 'does the series satisfy the necessary condition for convergence?', the answer is no, and the work to justify it is already done. Candidates who treat the L > 1 case as 'just divergence' leave points on the table.
The L = 1 case: the inconclusive verdict and how to switch methods
The L = 1 case is where AP Calculus candidates lose the most points per minute spent. The ratio test itself gives no information, and the rubric is unforgiving: stating 'diverges' or 'converges' on the basis of L = 1 scores zero on the conclusion row. The test writers expect the candidate to (a) compute L = 1, (b) explicitly state the test is inconclusive, and (c) switch to a different convergence test. The switch is itself part of the scored argument.
Worked example. Let aₙ = 1/n². Compute |aₙ₊₁ / aₙ| = (1/(n+1)²) · (n²/1) = n²/(n+1)². Take the limit: as n → ∞, n²/(n+1)² → 1, so L = 1. The candidate must write: 'L = 1, so the ratio test is inconclusive. The series is a p-series with p = 2 > 1, so the series converges.' That four-line argument scores the limit row, the inconclusive statement, and the convergence conclusion via a second test.
Three common errors on this case, all of which appear in the published scoring guidelines as zero-point answers:
- Computing L = 1 and then writing 'the series converges' with no further justification. The candidate has demonstrated the inconclusive case but has not actually shown convergence, so the conclusion point is not earned.
- Computing L = 1 and then writing 'the series diverges' by analogy with the harmonic series. The harmonic series is the wrong default; convergence depends on the specific aₙ, and the candidate must do the comparison work to claim a verdict.
- Switching methods silently — stating the answer is a p-series without writing the p-series comparison or the integral test. The rubric awards a setup point and a conclusion point for the new test; missing the setup costs one full point.
For most candidates, the L = 1 case is the difference between a 4 and a 5 on FRQs that involve series. Preparation should therefore weight this case heavily, with at least a third of the practice time spent on series where the ratio test fails to give a verdict. Limit comparison, direct comparison, and the integral test are the three switches the exam expects; the alternating series test is also valid if aₙ alternates in sign.
Common pitfalls and how to avoid them on the AP exam
Five pitfalls account for the majority of lost points on ratio-test items. Each has a specific counter-move that candidates can drill.
Pitfall 1: forgetting the absolute value in the limit. The ratio test is defined for lim |aₙ₊₁ / aₙ|. For positive-term series the bars are visually redundant, but the rubric sometimes awards a setup point only when the bars are present. Counter-move: write the absolute value bars in the first line of every ratio-test computation, even when the terms are obviously positive.
Pitfall 2: cancelling n's before substituting n+1. A common algebraic error is to write aₙ = 3ⁿ/n³, then substitute and immediately cancel the n's in the new expression, accidentally cancelling a₁ or a₀ instead of an aₙ factor. Counter-move: write the full expression for aₙ₊₁ before cancelling anything, and cancel one term at a time so the simplification is auditable.
Pitfall 3: stopping at the limit. Computing L is necessary but not sufficient. The conclusion row is independent. Counter-move: treat every ratio-test computation as a two-line argument — limit, then verdict — and never let the line break fall between them.
Pitfall 4: ignoring the inconclusive case. The L = 1 result is the most common ratio-test trap on the exam. Counter-move: when L = 1 lands on the page, write 'inconclusive' immediately and reach for a second test. Do not attempt to interpret L = 1 as 'probably converges' or 'probably diverges'.
Pitfall 5: running the ratio test on a series that does not need it. A p-series, a polynomial term, or a trigonometric term will produce L = 1 and force a method switch. Counter-move: spend five seconds scanning the term before starting the computation. If the dominant structure is a polynomial in n, choose comparison or p-series. If it is an exponential or factorial, choose the ratio test.
Comparing the ratio test with other AP Calculus convergence tools
The ratio test is one of roughly seven convergence tools on the AP Calculus BC syllabus. Knowing when each is cheapest is a scoring skill in its own right. The table below summarises the structural signature of aₙ that points to each test, the typical output, and whether the test gives additional information beyond a yes/no verdict.
| Test | Best structural fit for aₙ | Typical output | Extra information |
|---|---|---|---|
| nth term (divergence) test | Quick pre-check on any series | Divergence if aₙ does not tend to 0 | Necessary condition for convergence |
| Integral test | Positive, decreasing, continuous-approximating aₙ | Converges or diverges by integral value | Approximation of partial sums when used with remainder |
| Direct comparison | Series bounded above or below by a known p/geometric series | Converges or diverges by inequality | None |
| Limit comparison | Series with same dominant growth as a known series | Shares verdict with comparison series | None |
| Ratio test | Factorials, exponentials in n, binomial coefficients | Converges (L<1), diverges (L>1), inconclusive (L=1) | Failure of nth term test when L>1 |
| Root test | Series where aₙ is an nth power | Same verdict structure as ratio test | Same |
| Alternating series test | Series with (-1)ⁿ or (-1)ⁿ⁺¹ and decreasing magnitude | Converges by monotonicity to 0 | Error bound on partial sums |
The ratio test and the root test give the same verdict in essentially every case the AP exam asks about, but the ratio test is usually easier to compute because the algebra of cancellation is more transparent. The root test appears rarely on the exam and is almost always a backup choice for terms like aₙ = (n/(2n+1))ⁿ where the nth power structure is the dominant feature.
Preparation strategy: drilling the ratio test for a 5 on the AP exam
Three habits separate a candidate who scores the limit row from a candidate who scores all three rows on a ratio-test FRQ. The first is to write the limit line, the inequality L < 1 (or L > 1), and the conclusion as a single connected sentence, never as three independent notes. Rubrics reward connected arguments; the scoring reader looks for the words 'by the ratio test' on the same line as the inequality, and missing the connective phrase costs a row.
The second habit is to practice the L = 1 case in isolation. A surprising number of candidates can do L < 1 and L > 1 cleanly but freeze on L = 1 because the answer is 'do more work'. Drill at least eight L = 1 problems during AP exam preparation, each with a forced method switch to a p-series, a limit comparison, or the integral test. The switch should be automatic, not improvised.
The third habit is to recognise the L = 1 trap from the term structure before computing. If aₙ is a polynomial-over-polynomial expression in n, the ratio will collapse to 1 and the test will be inconclusive. Five seconds of inspection saves two minutes of wasted algebra. This is a method-selection skill, and it is the same skill that the MCQ section rewards in the back half of the calculator-permitted portion.
On the exam format itself: the BC exam contains 45 multiple choice questions in 105 minutes (about 2 minutes 20 seconds per question) and 6 free response questions in 90 minutes (15 minutes per question). Convergence and series FRQs typically appear as one of the six, sometimes paired with a Taylor polynomial sub-part. A ratio-test sub-part on such a question usually allows 4 to 6 minutes of work; faster than that signals that the L = 1 case has been missed or that the limit computation was skipped.
For scoring purposes, the ratio test on an FRQ typically contributes 2 to 3 points out of the 9 on the question. A candidate who scores all three rows on a 5-level FRQ is in strong shape for an overall 5; a candidate who scores only the limit row on every series FRQ typically lands at a 3 or low 4. The L = 1 case is the row that separates those outcomes, and it is the case to drill most intensively in the final six weeks of AP Calculus BC preparation.
Worked FRQ-style example: combining the ratio test with a method switch
Consider a series of the form ∑ (-1)ⁿ / (n² + 1). A candidate who reaches for the ratio test first will compute |aₙ₊₁ / aₙ| = (n² + 1) / ((n+1)² + 1). The numerator and denominator both behave like n² as n grows, so the limit is 1. The ratio test is inconclusive. A candidate who has practised the L = 1 trap will, at this point, switch to the alternating series test: the terms alternate in sign, the magnitudes 1/(n² + 1) decrease monotonically to 0, and the alternating series test therefore gives convergence. The argument on the page is: 'L = 1, so the ratio test is inconclusive. By the alternating series test, since 1/(n²+1) is positive, decreasing, and tends to 0, the series converges.' That is the full scored argument, and it earns all three rows the rubric is looking for.
A second worked example. Consider ∑ 2ⁿ/n!. The ratio test gives |aₙ₊₁ / aₙ| = 2ⁿ⁺¹/(n+1)! · n!/2ⁿ = 2/(n+1). The limit as n → ∞ is 0, which is less than 1, so the series converges absolutely. The argument is two lines and scores all three rows. Notice that this problem also connects to a Taylor series for e², but the ratio test does not require that connection; it is enough to write the convergence verdict.
A third worked example, on the L > 1 side. Consider ∑ n · 2ⁿ. The ratio is (n+1)·2ⁿ⁺¹ / (n · 2ⁿ) = 2(n+1)/n. The limit as n → ∞ is 2, which is greater than 1, so the series diverges. The bonus row — 'in particular, aₙ does not tend to 0' — is automatic, and the candidate can use it to answer a follow-up part about the divergence test without redoing any work.
These three examples, taken together, cover the three cases of the ratio test in roughly the proportions the AP exam asks them. Practice them until the limit computation, the inequality, and the conclusion are a single motion; that fluency is what the scoring guidelines are really measuring.
Conclusion and next steps
The ratio test on the AP Calculus BC exam is a structured argument with three scored rows: the limit itself, the inequality or inconclusive verdict, and the conclusion that follows from the limit. A 5-level answer treats those rows as a connected sentence rather than a sequence of independent notes, and a 5-level preparation plan weights the L = 1 inconclusive case at least as heavily as the L < 1 and L > 1 cases combined. The next concrete step is to drill eight ratio-test FRQs in timed conditions, with explicit attention to the switch from the ratio test to a second test whenever L = 1 appears, and to mark the rubric rows on each one to confirm full credit. AP Courses' AP Calculus BC series-and-convergence module walks students through exactly this drill, with rubric-aligned scoring on the L = 1 inconclusive case and the L > 1 bonus row that distinguish a 4 from a 5.