Sensitivity (Meta) Formula

Sensitivity (meta) is the degree to which a result or output changes in response to small changes in its inputs, parameters, or assumptions.

The Formula

sensitivity≈ΔoutputΔinput\text{sensitivity} \approx \frac{\Delta\text{output}}{\Delta\text{input}} (how much the output changes per unit change in input)

When to use: Is this result stable, or does a tiny change blow everything up?

Quick Example

Quadratic formula: small change in coefficients →\to small change in roots (usually). But near discriminant =0= 0, very sensitive.

Notation

Δ\Delta denotes a small change; high ΔoutputΔinput\frac{\Delta\text{output}}{\Delta\text{input}} means high sensitivity

What This Formula Means

The degree to which a result or output changes in response to small changes in its inputs, parameters, or assumptions.

Is this result stable, or does a tiny change blow everything up?

Formal View

The condition number κ=∣x⋅f′(x)f(x)∣\kappa = \left|\frac{x \cdot f'(x)}{f(x)}\right| measures relative sensitivity; κ≫1\kappa \gg 1 means the problem is ill-conditioned

Worked Examples

Example 1

easy
Compute f(x)=x3f(x) = x^3 at x=2x = 2 and x=2.1x = 2.1. Find the sensitivity: by what percentage does ff change when xx changes by 5%?

Answer

5% change in x⇒≈15.8% change in f\text{5\% change in }x \Rightarrow \approx 15.8\%\text{ change in }f

First step

1
f(2)=8f(2) = 8, f(2.1)=9.261f(2.1) = 9.261.

Full solution

  1. 2
    Change in ff: 9.261−8=1.2619.261 - 8 = 1.261. Relative change in ff: 1.2618≈15.8%\frac{1.261}{8} \approx 15.8\%.
  2. 3
    Change in xx: 0.12=5%\frac{0.1}{2} = 5\%.
  3. 4
    Sensitivity: a 5% increase in xx causes about a 15.8% increase in ff. The function is sensitive — it amplifies errors by a factor of about 3.
Sensitivity measures how much the output changes relative to the input. For f(x)=x3f(x) = x^3, the exponent 3 amplifies percentage changes roughly threefold, which can be derived from the derivative: f′(x)/f(x)≈3Δx/xf'(x)/f(x) \approx 3 \Delta x/x.

Example 2

medium
In the compound interest formula A=P(1+r)tA = P(1+r)^t, compute the sensitivity of AA to a small change Δr\Delta r in the interest rate, using the derivative dAdr\frac{dA}{dr}.

Example 3

medium
A retirement formula A=P(1+r)tA=P(1+r)^t with P=10000P=10000, t=20t=20 years. Compute AA for r=0.06r=0.06 vs r=0.07r=0.07 to gauge sensitivity to the rate.

Common Mistakes

  • Reporting the raw output change as sensitivity — divide by the input change to get the ratio.
  • Ignoring low-sensitivity inputs entirely — they're cheap to estimate, but identifying them is the point of the analysis.
  • Confusing sensitivity (the measure) with robustness (the property) — high sensitivity is what makes a result NOT robust.

Why This Formula Matters

In a model with many inputs, sensitivity tells you which one to measure most carefully and which won't matter; a result with huge sensitivity is fragile (small input error blows up), while low sensitivity signals robustness. It directs effort to the inputs that actually control the answer. Recognizing it by "Am I measuring how much the output moves PER unit change in the input?" — rather than by familiar numbers — is what lets a student tell it apart from robustness and slope / derivative and error propagation in a mixed problem set.

Frequently Asked Questions

What is the Sensitivity (Meta) formula?

The degree to which a result or output changes in response to small changes in its inputs, parameters, or assumptions.

How do you use the Sensitivity (Meta) formula?

Is this result stable, or does a tiny change blow everything up?

What do the symbols mean in the Sensitivity (Meta) formula?

Δ\Delta denotes a small change; high ΔoutputΔinput\frac{\Delta\text{output}}{\Delta\text{input}} means high sensitivity

Why is the Sensitivity (Meta) formula important in Math?

In a model with many inputs, sensitivity tells you which one to measure most carefully and which won't matter; a result with huge sensitivity is fragile (small input error blows up), while low sensitivity signals robustness. It directs effort to the inputs that actually control the answer. Recognizing it by "Am I measuring how much the output moves PER unit change in the input?" — rather than by familiar numbers — is what lets a student tell it apart from robustness and slope / derivative and error propagation in a mixed problem set.

What do students get wrong about Sensitivity (Meta)?

The procedure for sensitivity (meta) is the easy part; the trap is reporting the raw output change as sensitivity. Asking "Am I measuring how much the output moves PER unit change in the input?" first is what keeps a correct-looking calculation from being attached to the wrong concept.

What should I learn before the Sensitivity (Meta) formula?

Before studying the Sensitivity (Meta) formula, you should understand: local vs global behavior.

← Back to Sensitivity (Meta) See All Examples Practice Problems

Related Formulas

Robustness Formula