Skip to main content
United StatesCalculusSyllabus dot point

How do you bound the error of a Taylor polynomial approximation using the Lagrange error bound?

Topic 10.12 Lagrange Error Bound: bound the error of a Taylor polynomial approximation using the Lagrange form of the remainder (BC).

A focused answer to AP Calculus BC Topic 10.12, bounding the error of a Taylor polynomial approximation with the Lagrange form of the remainder, using a bound on the next derivative, with worked examples.

Generated by Claude Opus 4.89 min answer

Reviewed by: AI editorial process; not yet individually human-reviewed

Have a quick question? Jump to the Q&A page

Jump to a section
  1. What this topic is asking
  2. The bound
  3. Reading the bound as the next term
  4. Finding M, the real work
  5. A worked error bound
  6. When to use Lagrange versus the alternating bound

What this topic is asking

The College Board (Topic 10.12, BC only) gives the Lagrange error bound, the general way to bound how far a Taylor polynomial Pn(x)P_n(x) can be from the true function value f(x)f(x). Unlike the alternating-series bound, it works for any Taylor approximation, using a bound on the next derivative.

The bound

Reading the bound as the next term

The cleanest way to remember the Lagrange bound is that it is the next Taylor term in absolute value, with the unknown derivative f(n+1)(a)f^{(n+1)}(a) replaced by a guaranteed upper bound MM. The degree-nn polynomial's first missing term is f(n+1)(a)(n+1)!(xa)n+1\frac{f^{(n+1)}(a)}{(n+1)!}(x - a)^{n+1}; bounding f(n+1)|f^{(n+1)}| by MM over the interval turns this into M(n+1)!xan+1\frac{M}{(n+1)!}|x - a|^{n+1}. This is why every ingredient is "n+1n+1": the next derivative, the next factorial, the next power. Holding this picture in mind prevents the frequent slip of using nn instead of n+1n+1 somewhere.

Finding M, the real work

The harder step is choosing MM, an upper bound on the size of the (n+1)(n+1)-th derivative on the interval between the center and the point. For functions whose derivatives are bounded by a constant, this is easy: every derivative of sinx\sin x or cosx\cos x has absolute value at most 11, so M=1M = 1 always works. For exe^x, the derivative is exe^x, which is increasing, so on [0,c][0, c] the maximum is ece^c, and you bound it by a convenient number (as the worked exam question bounds e0.5e^{0.5} by 22). You do not need the exact maximum, only a valid upper bound, and a slightly generous MM still gives a correct (if looser) error bound. Justifying the choice of MM is what earns the reasoning marks.

A worked error bound

When to use Lagrange versus the alternating bound

The two error tools of Unit 10 cover different situations. The alternating series error bound (Topic 10.10) is simpler but applies only when the series you are truncating is alternating and satisfies the alternating series test; the bound is just the first omitted term. The Lagrange bound is general: it works for any Taylor polynomial, alternating or not, but requires you to bound a derivative. A good strategy is to use the alternating bound when the Taylor series happens to alternate at the point in question (often the case for sin\sin, cos\cos, ln(1+x)\ln(1+x)), and the Lagrange bound otherwise, or whenever the problem explicitly says "Lagrange error bound." Both give a valid (possibly different) bound; the alternating one is usually tighter and faster when it applies.

Exam-style practice questions

Practice questions written in the style of College Board exam questions on this dot point, with worked answer explainers. The year tag is the paper they imitate, not the source.

AP 2022 (BC, style)1 marksSection I (multiple choice, no calculator). The Lagrange error bound for a degree-nn Taylor approximation at xx about aa uses (A) M(n+1)!xan+1\frac{M}{(n+1)!}|x - a|^{n+1} (B) Mn!xan\frac{M}{n!}|x - a|^n (C) Mxan+1M|x - a|^{n+1} (D) M(n+1)!xan\frac{M}{(n+1)!}|x - a|^n
Show worked answer →

The correct answer is (A), M(n+1)!xan+1\frac{M}{(n+1)!}|x - a|^{n+1}.

If f(n+1)(t)M|f^{(n+1)}(t)| \le M between aa and xx, the remainder satisfies Rn(x)M(n+1)!xan+1|R_n(x)| \le \frac{M}{(n+1)!}|x - a|^{n+1}. It uses the (n+1)(n+1)-th derivative bound and the (n+1)(n+1)-th power.

AP 2024 (BC, style)4 marksSection II (free response, no calculator). Let P2(x)P_2(x) be the second-degree Maclaurin polynomial for f(x)=exf(x) = e^x. (a) Write P2(x)P_2(x). (b) Bound the error f(0.5)P2(0.5)|f(0.5) - P_2(0.5)| using the Lagrange bound with M=e0.5<2M = e^{0.5} < 2.
Show worked answer →

A 4-point Lagrange-bound problem.

(a) (1 point) P2(x)=1+x+x22P_2(x) = 1 + x + \frac{x^2}{2}.
(b) (3 points) The third derivative is exe^x, bounded by M=2M = 2 on [0,0.5][0, 0.5]. R2(0.5)M3!0.503=26(0.125)=0.2560.0417|R_2(0.5)| \le \frac{M}{3!}|0.5 - 0|^3 = \frac{2}{6}(0.125) = \frac{0.25}{6} \approx 0.0417.

Related dot points

Sources & how we know this