What Is Newton's Method?
Newton's method is a way to approximate a root of an equation. A root is an input value that makes a function equal to zero, so the goal is to find an \(x\) value where:
$$
f(x) = 0
$$
Many equations can be solved exactly, but many others cannot be solved neatly by algebra alone. Newton's method gives a practical numerical approach: start with a reasonable first guess, use the tangent line at that point to make a better guess, and repeat.
The method is also called the Newton-Raphson method. It is especially useful in calculus, numerical analysis, engineering, physics, economics, and other fields where nonlinear equations appear.
Why Newton's Method Matters
Newton's method matters because real problems often lead to equations that do not have a simple exact solution. A formula might be too complicated, unavailable, or impossible to rearrange into a clean expression for \(x\).
Instead of solving the equation symbolically, Newton's method builds a sequence of approximations:
$$
x_0, x_1, x_2, x_3, \ldots
$$
When the method works well, the sequence often moves toward the root very quickly. This makes Newton's method one of the most important root-finding tools in numerical mathematics.
Key Terms to Know
-
Root or zero: A value of \(x\) that makes \(f(x) = 0\).
-
Initial guess: The starting value, usually written as \(x_0\).
-
Iteration: One repeated step of the method.
-
Tangent line: A line that touches the curve at the current estimate and has the same slope as the curve there.
-
Derivative: The slope of the function at a point, written as \(f'(x)\).
-
Numerical derivative: An approximate derivative found from nearby function values instead of a symbolic derivative formula.
-
Residual: The value \(|f(x)|\) at the current estimate. A smaller residual means the estimate makes the function closer to zero.
-
Step error: The change between two successive estimates, such as \(|x_{n+1} - x_n|\).
-
Tolerance: The stopping threshold that says when the result is close enough for the chosen purpose.
-
Convergence: The estimates approach a useful root.
-
Divergence: The estimates move away, become unstable, or fail to settle near a root.
How Newton's Method Works
Newton's method is based on a simple geometric idea. Near a point on a smooth curve, the tangent line often gives a good local approximation to the curve. If the tangent line crosses the \(x\)-axis, that crossing point can be used as the next estimate of the root.
The standard Newton update is:
$$
x_{n+1} = x_n - \frac{f(x_n)}{f'(x_n)}
$$
Where:
-
\(x_n\) is the current estimate.
-
\(x_{n+1}\) is the next estimate.
-
\(f(x_n)\) is the function value at the current estimate.
-
\(f'(x_n)\) is the derivative, or slope, at the current estimate.
The formula comes from the tangent line approximation:
$$
y = f(x_n) + f'(x_n)(x - x_n)
$$
To find where this tangent line crosses the \(x\)-axis, set \(y = 0\) and solve for \(x\). That solution is the next estimate \(x_{n+1}\).
For supported parser syntax, this calculator obtains \(f'(x_n)\) from a safe analytic derivative rule. When that rule is unavailable, it compares centered finite differences at multiple step sizes and can use a second-order one-sided fallback near a domain boundary. A centered fallback estimate has the form:
$$
f'(x) \approx \frac{f(x+h) - f(x-h)}{2h}
$$
The table shows the parser-derived analytic derivative when available, or a numerical fallback estimate otherwise. If fallback samples cannot provide a sufficiently consistent estimate, the calculation stops instead of silently trusting the slope.
Examples of Newton's Method in Practice
Example 1: Approximating \(\sqrt{2}\)
To approximate \(\sqrt{2}\), solve:
$$
x^2 - 2 = 0
$$
Let:
$$
f(x) = x^2 - 2
$$
The derivative is:
$$
f'(x) = 2x
$$
Start with \(x_0 = 1.5\).
First iteration:
$$
x_1 = 1.5 - \frac{1.5^2 - 2}{2(1.5)}
$$
$$
x_1 = 1.5 - \frac{0.25}{3} \approx 1.416666667
$$
Second iteration:
$$
x_2 = 1.416666667 - \frac{1.416666667^2 - 2}{2(1.416666667)}
$$
$$
x_2 \approx 1.414215686
$$
The true value is approximately \(1.414213562\), so the estimate becomes accurate very quickly.
Example 2: Solving a Cubic Equation
Suppose a cube has volume \(20\) cubic units and you want its side length. The side length \(x\) satisfies:
$$
x^3 = 20
$$
Rewrite it as a root-finding problem:
$$
f(x) = x^3 - 20
$$
Newton's method uses:
$$
f'(x) = 3x^2
$$
Starting with \(x_0 = 3\):
$$
x_1 = 3 - \frac{3^3 - 20}{3(3^2)}
$$
$$
x_1 = 3 - \frac{7}{27} \approx 2.740740741
$$
A second iteration gives:
$$
x_2 = 2.7146696245795345 \approx 2.714669625
$$
The root is about \(2.714417617\), so after only two steps the estimate is already close.
Example 3: A Cycling Case
Newton's method does not always converge. For example, use:
$$
f(x) = x^3 - 2x + 2
$$
with initial guess \(x_0 = 0\).
Since:
$$
f(0) = 2
$$
and:
$$
f'(0) = -2
$$
Newton's update gives:
$$
x_1 = 0 - \frac{2}{-2} = 1
$$
At \(x_1 = 1\):
$$
f(1) = 1
$$
and:
$$
f'(1) = 1
$$
So:
$$
x_2 = 1 - \frac{1}{1} = 0
$$
The estimates alternate between \(0\) and \(1\) instead of moving toward a root. This is why convergence status and warning messages matter.
How to Interpret the Result
The root estimate is the calculator's current approximation of a value of \(x\) that makes \(f(x)\) equal to zero. A good estimate should usually have both a small residual and a small recent step change.
The residual \(|f(x)|\) tells you how close the function value is to zero at the final estimate. If the residual is small, the estimate nearly satisfies the equation.
The absolute step tells you how much the estimate changed in the most recent update:
$$
\text{absolute step} = |x_{n+1} - x_n|
$$
The relative step divides that change by \(\max(1,|x_n|,|x_{n+1}|)\). A small step alone does not prove that the estimate is a root, so the calculator also requires the normalized residual to pass.
The iteration count shows how many Newton updates were generated. Fewer iterations can indicate fast convergence, but a small iteration count is only meaningful when the status is good and the residual is acceptable.
The iteration table shows each step: the current estimate, function value, estimated derivative, next estimate, and error. This table is useful for checking whether the estimates are moving steadily toward a root or showing warning signs.
The graph is a visual aid. It can help you see the function curve, tangent steps, iteration points, and final root marker, but it should not be treated as a proof that the result is valid.
Common Mistakes and Misconceptions
A common mistake is assuming Newton's method always converges. It often works very well, but it can fail when the starting guess is poor, the derivative is near zero, the function is discontinuous, or the iteration falls into a cycle.
Another mistake is ignoring the initial guess. Newton's method is an open method, meaning it does not require two values that bracket a root. That can make it fast, but it also means the starting value strongly affects the result.
Users also sometimes trust the root estimate without checking the residual. The displayed \(x\) value may look stable, but \(|f(x)|\) tells you whether the function is actually close to zero.
For trigonometric functions, use radians. For example, \(\sin(x)\) treats \(x\) as a radian measure, not a degree measure.
Use a decimal point, not a comma. Enter \(0.5\), not \(0,5\).
Use the variable \(x\). Functions with other variables are not supported.
The derivative shown in the table comes from a safe parser-aware analytic rule when available. Numerical fallback values can be affected by rounding, very small step sizes, extreme function behavior, and domain restrictions.
When to Use Newton's Method
Use Newton's method when:
-
You need to approximate a solution to \(f(x) = 0\).
-
The function is reasonably smooth near the root.
-
You have a sensible initial guess.
-
You want a fast iterative method.
-
You can check the result with a residual, graph, table, or another method.
Newton's method is especially useful for nonlinear equations, square-root style equations, intersections that can be rewritten as \(f(x)=0\), and models where exact algebraic solving is inconvenient.
If you need guaranteed convergence over an interval, a bracketing method such as bisection may be safer. If you do not have a derivative or numerical derivative behavior is unstable, a secant method or other root-finding method may be more appropriate.
Limitations and Things to Keep in Mind
Newton's method is powerful, but it is not automatic proof of a root. The result depends on the function, the initial guess, the tolerance, and the behavior of the derivative.
Important limitations include:
-
Only real-valued, single-variable functions of \(x\) are supported.
-
Complex roots are not supported.
-
Users do not enter a separate symbolic derivative; supported parser syntax is differentiated analytically when safe, with adaptive finite differences as a fallback.
-
Numerical fallback estimates can be unreliable near discontinuities, sharp corners, noisy behavior, or domain boundaries.
-
If \(f'(x_n)\) is very close to zero, the Newton step can become extremely large or unstable.
-
Multiple nearby roots can make the method converge to a root other than the one you expected.
-
Poor initial guesses can cause divergence, cycling, or convergence to an unintended root.
-
The maximum iteration count is limited to \(80\) iterations.
-
Very small nonzero values are displayed in scientific notation rather than rounded to \(0\).
-
The graph range is chosen automatically and may not show every important feature of the function.
-
The graph may omit extreme values, so vertical asymptotes or very large function values may not be fully visible.
The calculator accepts a tolerance from \(10^{-14}\) through \(0.1\). Except when the function evaluates to exactly zero, convergence requires both the normalized residual and the relative step to meet that tolerance. This avoids accepting a merely small step or a scale-dependent residual as proof of convergence.
For homework, teaching, or exploratory math, Newton's method is a useful way to understand numerical root finding. For decisions involving money, safety, engineering design, official records, health, or legal obligations, verify the result independently and consult a qualified professional when appropriate.
How to Use This Calculator
-
Enter a supported expression for \(f(x)\) using \(x\) as the variable.
-
Enter a finite initial guess for \(x_0\).
-
Enter a tolerance from
1e-14 through 0.1.
-
Enter an integer maximum iteration count from \(1\) through \(80\).
-
Run the calculation, step through the method, or animate the steps using the available controls.
-
After a run, adjust the visible steps control to choose how many Newton steps appear on the graph.
-
Review the root estimate, status, absolute and normalized residuals, relative step, derivative quality, iteration table, formula explanation, and graph before relying on the result.
Supported constants include \(\pi\) and \(e\). Supported functions include sin, cos, tan, asin, acos, atan, sqrt, cbrt, abs, exp, ln, and log.
Frequently Asked Questions
Is Newton's method the same as the Newton-Raphson method?
Yes. In many textbooks and calculators, Newton's method and the Newton-Raphson method refer to the same iterative root-finding formula:
$$
x_{n+1} = x_n - \frac{f(x_n)}{f'(x_n)}
$$
Why does the initial guess matter so much?
Newton's method follows tangent lines from the current estimate. If the starting point is near the intended root and the function behaves smoothly, the method may converge quickly. If the starting point is far away, near a flat derivative, or on a troublesome part of the function, the method can move toward the wrong place or fail to converge.
What does tolerance mean?
Tolerance is the threshold used to decide when the approximation is close enough. In this calculator, convergence can be accepted when the residual \(|f(x)|\) is small enough or when the step change \(|x_{n+1} - x_n|\) is small enough. A smaller tolerance asks for a stricter result, but it may require more iterations or expose numerical limits.
Why might the calculator warn that the derivative is near zero?
Newton's formula divides by \(f'(x_n)\). If the derivative is very close to zero, the denominator is tiny, and the next step can become very large or unstable. A near-zero derivative warning means the displayed estimate should not be trusted without additional checking.
Can Newton's method find every root of a function?
No. A single Newton run starts from one initial guess and attempts to find one root. A function may have several roots, and different initial guesses can lead to different roots, cycling, divergence, or no useful result.
Why is my trigonometric result different from what I expected?
Trigonometric inputs use radians. If you enter \(\sin(30)\), the calculator treats \(30\) as radians, not \(30^\circ\). Convert degrees to radians first when needed:
$$
\text{radians} = \text{degrees} \times \frac{\pi}{180}
$$
Sources and References
Books and Open Textbooks
-
Gilbert Strang and Edwin “Jed” Herman. Calculus Volume 1. OpenStax, Section 4.9, “Newton’s Method.” Accessed July 4, 2026. OpenStax: Newton’s Method.
-
Autar Kaw. Numerical Methods with Applications. University of South Florida, Section 3.04, “Newton-Raphson Method for Solving a Nonlinear Equation.” Accessed July 4, 2026. LibreTexts: Newton-Raphson Method.
Online and Educational Sources
-
Carl Greco. “15.2: Numerical Differentiation.” Engineering Modeling and Analysis with Python, Arkansas Tech University / Engineering LibreTexts. Accessed July 4, 2026. https://eng.libretexts.org/Courses/ArkansasTechUniversity/EngineeringModelingandAnalysiswithPython/15%3ADynamicSystems/15.02%3ANumerical_Differentiation
