Approximating Derivatives Numerically

What Numerical Differentiation Is For

Sometimes you are asked for a derivative even though you are not given an equation that you can differentiate directly. Instead, you may be given a table of values, a graph, or measurements from an experiment. In these situations, you can approximate the derivative numerically.

The key idea is simple: a derivative is an instantaneous rate of change, while a secant slope measures the average rate of change between two nearby points. If those points are sufficiently close, the secant slope can provide a good approximation to the derivative.

As the second point gets closer and closer to the point of interest, the secant slope approaches the tangent slope. Numerical differentiation uses nearby values to estimate that tangent slope.

The Basic Difference Quotient

Suppose you want to approximate . If you know the function values at and , you can use the slope between those two points:

This is a forward difference because the second point is to the right of . It estimates the derivative using the current point and a nearby point to its right.

Forward Difference

The forward difference formula is especially useful when a table gives you the value at and the next value after it:

For a table of values, the most important thing is to use the actual spacing between the -values. Do not automatically divide by 1 unless the -values are actually one unit apart.

Backward Difference

You can also estimate a derivative using a point immediately to the left of the point of interest. This is called a backward difference.

Notice that this is still just a slope formula. The numerator is the change in the function values, and the denominator is the corresponding change in the input values.

Symmetric Difference: Often the Best Approximation

If you have values on both sides of the point where you want the derivative, a symmetric difference can often give a better approximation than using only one side.

Here, the points are equally spaced around . The approximation uses the slope between the two outer points. Because the points are balanced on either side of the target, this method often reduces the bias that can occur when using only a forward or backward difference.

Why the Symmetric Difference Works

It helps to connect the numerical formula back to the definition of the derivative. The derivative at is the limit of the difference quotient as the interval becomes arbitrarily small:

A symmetric approximation instead uses a small distance on both sides:

As becomes small, this gives a close estimate of the tangent slope at . On AP Calculus problems, you generally do not need to prove why the approximation works. You need to recognize the appropriate difference quotient and calculate it correctly.

Choosing the Right Formula

• Forward difference: use the point of interest and a nearby point to the right.
• Backward difference: use the point of interest and a nearby point to the left.
• Symmetric difference: use equally spaced points on both sides when they are available.
• Secant slope: whenever you are estimating a derivative from two nearby points, remember that you are approximating the tangent slope with a secant slope.

Using a Graph

Numerical differentiation is not limited to tables. If a graph gives you coordinates of two points near , you can estimate the derivative by finding the slope between them.

For example, if two nearby points on the curve are and , then:

The closer the points are to the target and to each other, the more useful the secant slope is as an approximation to the tangent slope—provided the data are sufficiently accurate.

Using a Table with Unequal Spacing

This is a common place to lose points. The denominator must use the actual difference between the -values.

Common Mistakes

Practice Problems