AI Tutor
Imagine running an algorithm designed to pinpoint a root in milliseconds, only for your calculator to throw an error or bounce back and forth between two numbers forever.
First developed by Isaac Newton in 1669 and refined by Joseph Raphson in 1690, the Newton-Raphson method uses tangent slopes to iteratively home in on roots using the formula xn+1=xn−f(xn)f′(xn)x_{n+1} = x_n - \frac{f(x_n)}{f'(x_n)}.
What is the most immediate way this algebraic formula can break down on paper?

1. Stationary Points and Zero Derivatives

A stationary point is any point on a curve where the derivative equals zero (f′(xn)=0f'(x_n) = 0), such as a local maximum, local minimum, or horizontal inflection point.
Because the iteration formula divides by f′(xn)f'(x_n), encountering a stationary point causes an undefined division by zero.
For example, to find a root of f(x)=x3−3x+1f(x) = x^3 - 3x + 1, the derivative is f′(x)=3x2−3f'(x) = 3x^2 - 3. If you choose the initial guess x0=1x_0 = 1, then f′(1)=3(1)2−3=0f'(1) = 3(1)^2 - 3 = 0, causing the algorithm to crash instantly.
What happens if the derivative is non-zero, but the geometry of the curve traps the tangent lines in an infinite loop?

2. Periodic Cycling

Cycling occurs when an initial value x0x_0 generates an estimate x1x_1, whose tangent line points right back to x0x_0.