Math · Calculus
The derivative as a rate of change
Try this first
A car's odometer reads 0 km at noon and 100 km at 1pm. So its average speed was 100 km/h. But the speedometer at 12:30 read 130 km/h. How can the speed at one instant differ from the average — and what would you measure to get the 12:30 number exactly?
Hold that car in mind. Over the whole hour you can only get an average rate: distance gone divided by time taken, 100 km ÷ 1 h. But to recover the speedometer reading at 12:30 you'd zoom in — measure the distance covered in the minute around 12:30, then the second around it, then the millisecond. As the window shrinks, that average rate settles on one number: the speed at that instant. That settling-down is the entire idea of a derivative.
The one idea
The derivative is the instantaneous rate of change: the average rate of change measured over an interval, then squeezed until the interval shrinks to zero. On a graph, that's the slope of the line that just grazes the curve at a point — the tangent.
From average to instantaneous
An average rate of change is just rise over run between two points on a curve, (f(b) − f(a)) / (b − a). Geometrically it's the slope of the secant line cutting through both points. Now slide the second point toward the first. The secant pivots, and in the limit — when the gap is infinitesimally small — it becomes the tangent line. Its slope is the derivative, written f′(x) or df/dx.
Formally, with a gap of size h that we send to zero:
f′(x) = limh→0 [ f(x+h) − f(x) ] / h
Notice the secant through Q is steeper than the tangent here, because the curve is bending upward — the average over the wide gap [2, 3.5] overstates the rate right at P. Shrinking the gap fixes that.
The same idea, three names
"Rate of change", "slope", and "derivative" are one concept seen from three angles. Knowing they're the same is most of the battle.
| Setting | What you call the rate | What it means |
|---|---|---|
| Physics | Velocity | How fast position changes per second |
| Geometry | Slope of the tangent | Steepness of the curve at a point |
| Calculus | Derivative f′(x) | Limit of average rates as the interval → 0 |
| Economics | Marginal cost | Cost added by one more unit |
What the sign and size tell you
The derivative is itself a number that varies from point to point, and reading it is half its value. Positive means the quantity is rising; negative means falling; zero means momentarily flat — the top of a hill, the bottom of a valley. A large magnitude means steep, fast change; a small one means gentle. The derivative is your dashboard for how something is moving, not just where it is.
Work one, then finish one
Worked: Take f(t) = t² (distance in metres, t in seconds). The average speed from t = 2 to t = 2 + h is [(2+h)² − 2²]/h = [4 + 4h + h² − 4]/h = (4h + h²)/h = 4 + h. As h → 0, this approaches 4 m/s. So the instantaneous speed at t = 2 is 4 — matching the rule f′(t) = 2t giving 2·2 = 4.
Your turn: Using the same trick on f(t) = t², find the instantaneous rate at t = 3. Expand [(3+h)² − 9]/h, simplify, and let h → 0. (Answer: 6 + h → 6, agreeing with 2·3.)
Why this earns a place in your toolkit
Rates of change are how machines learn. Training a model means repeatedly asking "if I nudge this weight, does the error go up or down, and how fast?" — that nudge-sensitivity is a derivative, and gradient descent is nothing but rolling downhill along those slopes. The same idea reads a curve's steepness in any dashboard: velocity from a position log, growth rate from a user-count chart, the marginal cost of one more API call. Once you see "rate of change", you see derivatives everywhere data moves.
Recall check · no peeking
- What is the geometric difference between a secant line and a tangent line, and how do you get from one to the other?
- Write the limit definition of
f′(x)from memory. What does theh → 0part accomplish? - If a derivative is negative and large in magnitude, what is the quantity doing?
- Why can the average speed over an hour differ from the speedometer reading at one instant?
Explain it back
In one plain sentence, tell a friend why "the derivative" and "the slope of the curve at a point" are the same thing.