Approximate calculations
Near x₀ the graph can be replaced by its tangent, so for small Δx we have f(x₀ + Δx) ≈ f(x₀) + f′(x₀) · Δx. To use it, choose x₀ so that f(x₀) and f′(x₀) are easy (perfect squares, cubes, 0, 1, 30°, 45°, …) and Δx small: the smaller Δx is, the better the approximation. At x₀ = 0: sin x ≈ x, tan x ≈ x (x in radians), eˣ ≈ 1 + x, ln(1 + x) ≈ x, (1 + x)ᵐ ≈ 1 + mx, √(1 + x) ≈ 1 + x/2. Do not forget to convert angles to radians: 1° = π/180 ≈ 0.01745 rad. Write results with ≈ to show they are approximate; the method gives a value to several decimal places even without a calculator.
A village’s population grows by 2 % a year. After 5 years it is about 1 + 5·0.02 = 1.10 times as large, i.e. 10 % more; the exact calculation 1.02⁵ ≈ 1.104 gives about the same. Without a calculator you get an answer accurate to 0.4 %, because for a small change the curve can be replaced by its tangent.
Engineers quickly estimate how much a measurement error affects a result, physicists replace sin x by x for small angles, and calculator and software developers compute values by approximation.
On paper, estimate 1.03³ with the shortcut (1 + 3·0.03), then check it by multiplying 1.03 by itself three times. Write down how big the difference between the two answers is.
“No calculator”: the class splits into two groups. One computes 5 expressions given by the teacher (say √9.06, 0.98⁴⁰, sin 0.05) with the small-increment formula, the other with a calculator. Compare the results and discuss when the error is larger.