Alternating Series Error Bound

SECTION 1

Historical Context & Motivation

Long before digital computers could evaluate sums to arbitrary precision, mathematicians faced a deeply practical question: if you must stop adding terms at some point, how far off is your partial sum from the true value of an infinite series? This question drove the development of error estimation techniques, which remain central to numerical analysis, physics, and engineering. The story of the alternating series error bound is intertwined with the broader effort to put infinite series on rigorous footing—a project that spanned centuries and involved some of the greatest minds in mathematics.

1714

Leibniz and the Alternating Harmonic Series

Gottfried Wilhelm Leibniz studied the series 1 − 1/2 + 1/3 − 1/4 + ⋯ and argued it converges to ln 2. He observed that successive partial sums oscillate around the limit, an insight foundational to the alternating series test.

1821

Cauchy's Cours d'analyse

Augustin-Louis Cauchy formalized convergence criteria for series, including a rigorous proof of the alternating series test that Leibniz had only heuristically described. His work established the framework for bounding truncation errors.

1837

Dirichlet's Contributions

Peter Gustav Lejeune Dirichlet refined the theory of conditionally convergent series and clarified when rearrangement of terms changes the sum—highlighting the special, well-behaved nature of alternating series whose terms decrease in absolute value.

1854

Riemann's Rearrangement Theorem

Bernhard Riemann proved that any conditionally convergent series can be rearranged to converge to any real number. This dramatic result underscored the importance of the alternating series structure, where the error bound provides a reliable guarantee precisely because the original term order is preserved.

The central question that the alternating series error bound addresses is this: when you truncate a convergent alternating series after n terms, what is the maximum possible error in your approximation? The answer turns out to be beautifully simple and is one of the most elegant results in the study of infinite series.

SECTION 2

Core Principles & Definitions

Before applying the error bound, you must confirm that the series satisfies the conditions of the Alternating Series Test (Leibniz Test). This test guarantees convergence and simultaneously provides the machinery for the error bound. Recall that an alternating series has the form Σ(−1)ⁿbₙ or Σ(−1)ⁿ⁺¹bₙ where bₙ > 0 for all n. The following foundational ideas make the error bound work.

Alternating Structure

Terms alternate in sign: each positive term is followed by a negative term and vice versa. This sign-switching is what causes partial sums to oscillate around the limit S.

Monotone Decrease

The absolute values of the terms must be non-increasing: bₙ₊₁ ≤ bₙ for all n beyond some index N. Each successive correction overshoots by less than the previous one.

Limit to Zero

The nth term must approach zero: lim(n→∞) bₙ = 0. This ensures that the oscillations shrink, forcing the partial sums to converge rather than perpetually bounce.

Error Bound Guarantee

Once both conditions are satisfied, the error |S − Sₙ| is bounded above by bₙ₊₁, the absolute value of the first omitted term. This is the alternating series error bound (ASEB).

✦ KEY TAKEAWAY

Think of a convergent alternating series like tightening a bolt by turning it alternately right and left, with each turn smaller than the last. Each partial sum overshoots the target in the opposite direction, but by a smaller amount. The maximum possible overshoot at any stage is exactly the size of the next turn you haven't yet made—that is, the first omitted term.

SECTION 3

Visual Explanation

The diagram below illustrates how partial sums of a convergent alternating series oscillate around the true sum S. Notice that S₁ overshoots above S, then S₂ corrects below S, then S₃ overshoots above again—but each overshoot is smaller than the last. The distance from any partial sum Sₙ to S is always less than the magnitude of the next term.

Partial sums S₁, S₂, S₃, … alternate above and below the true sum S (dashed cyan line). The vertical dashed segments show that each error |S − Sₙ| is bounded by the magnitude of the next term bₙ₊₁.

The oscillation pattern is the geometric heart of the error bound. Because the partial sums alternate above and below S, and each step is smaller than the previous one, the true sum S is always trapped between any two consecutive partial sums. This trapping property means that |S − Sₙ| ≤ |Sₙ₊₁ − Sₙ| = bₙ₊₁, which is precisely the alternating series error bound.

SECTION 4

Mathematical Framework

Let us formalize the result. Consider an alternating series that satisfies the conditions of the Alternating Series Test, and let S denote its sum and Sₙ the nth partial sum.

ALTERNATING SERIES FORM

S = Σₙ₌₁∞ (−1)ⁿ⁺¹ bₙ = b₁ − b₂ + b₃ − b₄ + ⋯

where bₙ > 0, bₙ₊₁ ≤ bₙ for all n, and lim(n→∞) bₙ = 0.

ALTERNATING SERIES ERROR BOUND (ASEB)

|S − Sₙ| ≤ bₙ₊₁

The absolute error in approximating S by the nth partial sum Sₙ is at most bₙ₊₁, the absolute value of the first omitted term.

Proof Sketch

The proof proceeds by examining the remainder Rₙ = S − Sₙ. Group the remaining terms in consecutive pairs: Rₙ = (−1)ⁿ⁺²bₙ₊₁ + (−1)ⁿ⁺³bₙ₊₂ + ⋯. Because bₙ₊₁ ≥ bₙ₊₂ ≥ bₙ₊₃ ≥ ⋯, each consecutive pair (bₖ − bₖ₊₁) ≥ 0. Therefore, the remainder has the same sign as its leading term and its magnitude is at most bₙ₊₁. More precisely, grouping the tail as (bₙ₊₁ − bₙ₊₂) + (bₙ₊₃ − bₙ₊₄) + ⋯ shows that |Rₙ| ≤ bₙ₊₁.

FINDING N FOR DESIRED ACCURACY

|S − Sₙ| < ε ⟸ bₙ₊₁ < ε

To guarantee accuracy within ε, find the smallest n such that the (n+1)th term's absolute value is less than ε. This determines how many terms you need to sum.

📝 AP Exam Tip

On the AP Calculus BC exam, the ASEB is frequently tested in free-response questions involving Taylor or Maclaurin series. You will be asked to approximate a function value and then bound the error. Always verify that the series is alternating and that the terms decrease in absolute value before applying the bound. If you forget to verify these conditions, you will lose justification points.

SECTION 5

Conditions & Common Pitfalls

The alternating series error bound is powerful, but it applies only under specific conditions. Misapplying it to series that do not meet these conditions is one of the most common errors students make on the AP exam. The following diagram provides a decision flowchart, and the table below catalogs typical pitfalls.

Flowchart for determining whether the Alternating Series Error Bound can be applied. If any condition fails, the ASEB cannot be used.

Common Pitfalls When Applying the Alternating Series Error Bound

Pitfall	Why It's Wrong	Correction
Using ASEB on a non-alternating series	The oscillation property does not hold; partial sums may approach S from one side only.	Use the Lagrange error bound or integral remainder estimate instead.
Using bₙ instead of bₙ₊₁	The error bound is the first omitted term, not the last included term.	If Sₙ is your approximation, the bound is the (n+1)th term: bₙ₊₁.
Forgetting to verify that terms decrease	If bₙ₊₁ > bₙ for some n, the trapping argument fails.	Show bₙ₊₁ ≤ bₙ (often via the first derivative test on the continuous analog).
Confusing the ASEB with Lagrange Error Bound	The Lagrange bound applies to Taylor polynomial remainders in general; the ASEB applies specifically when the Taylor series is alternating.	When a Taylor series is alternating and satisfies AST conditions, the ASEB is often simpler. Choose the tool that fits.

SECTION 6

Worked Example

Let us approximate cos(1) using the Maclaurin series for cos x and bound the error using the ASEB.

Approximate cos(1) with |error| < 0.001

Step 1 — Write the Maclaurin SeriesThe Maclaurin series for cos x is cos x = Σₙ₌₀∞ (−1)ⁿ x²ⁿ/(2n)! = 1 − x²/2! + x⁴/4! − x⁶/6! + ⋯. At x = 1, we get cos(1) = 1 − 1/2! + 1/4! − 1/6! + ⋯.

Step 2 — Verify ASEB ConditionsThe series alternates in sign: ✓. Define bₙ = 1/(2n)!. Since (2n+2)! = (2n+2)(2n+1)(2n)!, we have bₙ₊₁ = 1/(2n+2)! < 1/(2n)! = bₙ, so terms decrease: ✓. And lim(n→∞) 1/(2n)! = 0: ✓. All three conditions are satisfied, so the ASEB applies.

Step 3 — Determine How Many Terms Are NeededWe need bₙ₊₁ < 0.001. Compute successive terms: b₀ = 1, b₁ = 1/2 = 0.5, b₂ = 1/24 ≈ 0.04167, b₃ = 1/720 ≈ 0.00139, b₄ = 1/40320 ≈ 0.0000248. Since b₄ ≈ 0.0000248 < 0.001, using the partial sum through b₃ (i.e., summing through the n = 3 term) gives an error less than 0.001. But note b₃ ≈ 0.00139 > 0.001, so stopping at b₂ would not suffice.

We need at least the first 4 nonzero terms (n = 0 through n = 3).

Step 4 — Compute the Partial SumS₄ = 1 − 1/2 + 1/24 − 1/720 = 1 − 0.5 + 0.04167 − 0.00139 = 0.54028 (to five decimal places).

cos(1) ≈ 0.54028

Step 5 — State the Error BoundBy the ASEB, |cos(1) − 0.54028| ≤ b₄ = 1/40320 ≈ 0.0000248. This is well within our target of 0.001. The actual value of cos(1) ≈ 0.54030, confirming our approximation is accurate to the claimed precision.

|error| ≤ 1/40320 ≈ 0.0000248 < 0.001 ✓

SECTION 7

ASEB vs. Other Error Bounds

The AP Calculus BC exam expects you to choose the appropriate error estimation technique based on the structure of the series or polynomial. The ASEB is the simplest of the major bounds, but it applies only in specific situations. Understanding when to use each tool is critical for both multiple-choice efficiency and free-response justifications.

Comparison of Error Bound Techniques on the AP Calculus BC Exam

Feature	Alternating Series Error Bound	Lagrange Error Bound
Applies to	Alternating series satisfying AST conditions	Any Taylor polynomial approximation with bounded (n+1)th derivative
Formula	\|S − Sₙ\| ≤ bₙ₊₁	\|Rₙ(x)\| ≤ M\|x − c\|ⁿ⁺¹/(n+1)!
Ease of use	Very simple: just evaluate the next term	Requires finding a bound M on the (n+1)th derivative
Tightness	Often very tight—actual error is frequently close to the bound	Can be a significant overestimate if M is loose
When both apply	Preferred when applicable—simpler and typically tighter	Use when the series is not alternating, or when you need a bound for a general Taylor polynomial

✦ KEY TAKEAWAY

Think of the ASEB and the Lagrange Error Bound as two different measuring tapes. The ASEB is like a precision digital caliper—simple, direct, and highly accurate—but it only works on alternating-series-shaped objects. The Lagrange bound is like a flexible tape measure—it works on anything, but may give you a looser measurement. On the AP exam, always use the ASEB when its conditions are met, because it is faster and typically yields a tighter bound.

SECTION 8

Connection to Taylor Series & Advanced Applications

The ASEB finds its most common application on the AP exam in the context of Taylor and Maclaurin series. Many well-known functions produce alternating Taylor series when evaluated at particular values of x. For example, sin x, cos x, eˣ (at negative x), and ln(1 + x) all yield alternating series under appropriate substitutions. When a Taylor series evaluated at a specific x-value produces an alternating series with decreasing terms, the ASEB provides a cleaner bound than the Lagrange remainder, which requires finding the maximum of the (n+1)th derivative on an interval.

Common Maclaurin Series That Yield Alternating Series

Function	Series (at given x)	Alternating for…
sin x	x − x³/3! + x⁵/5! − ⋯	All x > 0 (after enough terms)
cos x	1 − x²/2! + x⁴/4! − ⋯	All x ≠ 0 (intrinsically alternating)
eˣ at x < 0	1 + x + x²/2! + ⋯ (alternates when x < 0)	x < 0 (e.g., e⁻¹)
ln(1 + x)	x − x²/2 + x³/3 − ⋯	0 < x ≤ 1

Beyond the AP exam, the ASEB appears in numerical analysis wherever alternating series arise—for instance, in evaluating special functions, computing definite integrals represented by power series, and in signal processing applications involving Fourier series with specific symmetry properties. The principle of the error being bounded by the first omitted term extends conceptually to other mathematical structures, such as asymptotic expansions in advanced mathematics, where similar trapping arguments apply even when the series does not converge in the classical sense.

SECTION 9

Practice Problems

The series Σₙ₌₁∞ (−1)ⁿ⁺¹ / n converges to ln 2. If you approximate ln 2 using the partial sum S₅ = 1 − 1/2 + 1/3 − 1/4 + 1/5, which of the following is a correct statement about the error?

What is the minimum number of nonzero terms of the Maclaurin series for sin x needed to approximate sin(1) with an error less than 0.001?

The function f is defined by f(x) = Σₙ₌₀∞ (−1)ⁿ xⁿ/(n² + 1). If T₃(x) denotes the third-degree Taylor polynomial for f about x = 0, what is the maximum error in using T₃(0.5) to approximate f(0.5)?

PROBLEM 4 — APPLIED

A function g has a Maclaurin series given by g(x) = Σₙ₌₀∞ (−1)ⁿ x²ⁿ⁺¹ / ((2n+1) · 3ⁿ). (a) Show that the series converges for x = 1. (b) Use the fourth-degree Maclaurin polynomial P₄(x) to approximate g(1). (c) Use the alternating series error bound to determine an upper bound for |g(1) − P₄(1)|. (d) Could the approximation P₄(1) be an overestimate or an underestimate of g(1)? Justify your answer.

PROBLEM 5 — CRITICAL THINKING

Consider the series S = Σₙ₌₁∞ (−1)ⁿ⁺¹ · n / (n² + 1). (a) Does this series converge? Justify. (b) A student claims that the ASEB gives |S − S₁₀| ≤ 11/122. Is the student's claim valid? Explain. (c) Without computing S exactly, determine whether S₁₀ is an overestimate or underestimate of S.

SUMMARY

Summary

The Alternating Series Error Bound states that for a convergent alternating series satisfying the conditions of the Alternating Series Test—alternating signs, decreasing terms in absolute value, and terms approaching zero—the error in using the nth partial sum Sₙ to approximate the true sum S is at most |S − Sₙ| ≤ bₙ₊₁, the absolute value of the first omitted term. This bound arises from the oscillation pattern of partial sums around the limit.

On the AP Calculus BC exam, the ASEB is most commonly applied to Taylor and Maclaurin series for functions like sin x, cos x, eˣ, and ln(1 + x). When applicable, it is simpler and often tighter than the Lagrange Error Bound. Always verify the three AST conditions before invoking the ASEB, identify whether the partial sum is an overestimate or underestimate based on the sign of the first omitted term, and remember that the bound uses bₙ₊₁ (the first term you did not include), not bₙ.