Secant Method — Definition, Formula & Examples
The secant method is a numerical technique for finding roots of an equation by repeatedly drawing secant lines through two recent approximations and using the x-intercept as the next guess.
Given two initial approximations and , the secant method generates a sequence defined by the recurrence , which converges to a root of under suitable conditions. It can be viewed as a finite-difference approximation to Newton's method, replacing the derivative with the slope of the secant line through and .
Key Formula
Where:
- = Current approximation to the root
- = Previous approximation to the root
- = Value of the function at the current approximation
- = Value of the function at the previous approximation
How It Works
Start with two initial guesses and near the root. Evaluate at both points and compute the slope of the secant line connecting them. Find where this secant line crosses the x-axis — that x-intercept becomes your next approximation . Then repeat the process using and , and so on. Stop when or is smaller than your chosen tolerance. Unlike Newton's method, you never need to compute or know the derivative of .
Worked Example
Problem: Use the secant method with and to find one iteration toward a root of .
Evaluate f at both points: Compute and .
Apply the secant formula: Substitute into the recurrence relation to find .
Check the result: Evaluate at the new approximation.
Answer: After one iteration, with , which is much closer to the true root than either starting guess.
Why It Matters
The secant method is widely used in engineering and scientific computing when a derivative is expensive or impossible to compute analytically. It converges faster than the bisection method (its order of convergence is approximately 1.618) while avoiding the derivative calculations that Newton's method requires. Courses in numerical analysis and computational mathematics treat it as a standard tool for nonlinear equation solving.
Common Mistakes
Mistake: Choosing two initial guesses where , which causes division by a near-zero denominator.
Correction: Pick starting values that give noticeably different function values so the secant line has a well-defined, non-extreme slope.
