AP® Calculus BC · Unit 10

What Is the Lagrange Error Bound in AP Calculus BC?

The Lagrange error bound puts a guaranteed ceiling on how far a Taylor polynomial can be from the true function value: the error is at most MM divided by (n+1)!(n+1)!, times ∣x−a∣n+1|x-a|^{n+1}, where MM is the largest size of the (n+1)(n+1)th derivative between the center aa and the point xx. It never gives the exact error, only a guaranteed limit.

The Lagrange error bound

If Pn(x)P_n(x) is the degree-nn Taylor polynomial for ff centered at aa, then ∣f(x)−Pn(x)∣≤M(n+1)!∣x−a∣n+1,|f(x) - P_n(x)| \le \frac{M}{(n+1)!}|x-a|^{n+1}, where MM is the maximum value of ∣f(n+1)(c)∣\left|f^{(n+1)}(c)\right| for cc between aa and xx — not at a single point, but the largest value anywhere on that whole interval.

Step-by-Step Method

  1. Identify nn (the degree of the Taylor polynomial you are using) and the interval between the center aa and the point xx.
  2. Find the (n+1)(n+1)th derivative of ff symbolically.
  3. Find MM, the maximum of ∣f(n+1)(c)∣\left|f^{(n+1)}(c)\right| for cc on the whole interval between aa and xx — check where that derivative is largest, not just its value at the endpoints.
  4. Plug MM, nn, and ∣x−a∣|x - a| into the bound formula.
  5. State the result as an inequality — the bound guarantees the error is at most that number, not that it equals it.

Worked Examples

Foundational — bounding the error in approximating cos(0.3)Use the Lagrange error bound to bound the error when the degree-2 Maclaurin polynomial for cos⁡x\cos x is used to approximate cos⁡(0.3)\cos(0.3).
  1. The degree-2 Maclaurin polynomial is P2(x)=1−x22P_2(x) = 1 - \dfrac{x^2}{2}, so P2(0.3)=1−0.092=0.955P_2(0.3) = 1 - \dfrac{0.09}{2} = 0.955.
  2. The remainder uses the 3rd derivative: f(x)=cos⁡x⇒f′′′(x)=sin⁡xf(x)=\cos x \Rightarrow f'''(x) = \sin x. Since sin⁡t\sin t is increasing on [0,0.3][0, 0.3], its maximum there is at t=0.3t = 0.3: M=sin⁡(0.3)≈0.2955M = \sin(0.3) \approx 0.2955.
  3. Bound: ∣R2(0.3)∣≤M3!(0.3)3=sin⁡(0.3)6(0.027)≈0.00133|R_2(0.3)| \le \dfrac{M}{3!}(0.3)^3 = \dfrac{\sin(0.3)}{6}(0.027) \approx 0.00133.
  4. Check: the actual value cos⁡(0.3)≈0.95534\cos(0.3) \approx 0.95534, so the true error is about 0.000340.00034 — well under the guaranteed bound of 0.001330.00133, exactly as the theorem promises.
Harder — how many terms to guarantee an error under 0.0001How many terms of the Maclaurin series for exe^x are needed to approximate e0.5e^{0.5} with error less than 0.00010.0001?
  1. For f(x)=exf(x) = e^x, every derivative is exe^x, so f(n+1)(t)=etf^{(n+1)}(t) = e^t. On [0,0.5][0, 0.5], ete^t is increasing, so its maximum is M=e0.5≈1.6487M = e^{0.5} \approx 1.6487.
  2. Try n=4n=4 (5 terms, through x4x^4): bound =e0.5(0.5)55!≈0.00043= \dfrac{e^{0.5}(0.5)^5}{5!} \approx 0.00043. That is bigger than 0.00010.0001, so 5 terms is not enough.
  3. Try n=5n=5 (6 terms, through x5x^5): bound =e0.5(0.5)66!≈0.0000358= \dfrac{e^{0.5}(0.5)^6}{6!} \approx 0.0000358. That is less than 0.00010.0001.
  4. So the degree-5 Taylor polynomial — 6 terms total — is the smallest one that guarantees error under 0.00010.0001.

Common Mistakes on the AP Exam

How This Appears on the AP Exam

Try a timed AP-style question on this topic — free.

Start Free

Frequently Asked Questions

What does the Lagrange error bound actually tell you?

It gives a guaranteed upper limit on how far a Taylor polynomial approximation can be from the true function value — not the exact error, just a number the true error is never allowed to exceed.

How do you find M in the Lagrange error bound formula?

Take the (n+1)(n+1)th derivative of the function, then find its largest absolute value anywhere on the interval between the center and the point you are approximating — not just at one endpoint.

Why does the bound use the (n+1)th derivative and not the nth?

Because the error is what is left over after the degree-nn polynomial has already matched the function’s value and its first nn derivatives at the center; the next derivative, the (n+1)(n+1)th, is what controls everything the polynomial could not capture.

Can the Lagrange error bound be larger than the actual error?

Yes, and usually is. It is a guaranteed ceiling, not an exact value, so the true error is typically smaller than the bound — sometimes much smaller, as in the cos(0.3) example on this page.

Is the Lagrange error bound tested on the AP Calculus AB exam?

No. It is part of Unit 10, the infinite series unit, which is tested only on the BC exam.

Related Topics

Keep going with timed tests & full mocks, study materials for every unit, DME Nova — your AI tutorthat won't make up an answer, and FRQ feedback scored against an exam-style rubric.

Start Free

Written by the DME AP Calculus team — 20+ years of AP & STEM teaching.

Last updated: 2026-10-08

AP® is a registered trademark of the College Board, which is not affiliated with, and does not endorse, this site.