# Double pendulum > How fast does a double pendulum turn chaotic? Two nearly identical double pendulums start a hair apart. Watch them swing together, then drift into completely different paths. Simulate two nearly identical double pendulums and see how a tiny difference grows into chaos. Drag the start angles, run it and read the drift time. - Page: https://how.box/double-pendulum/ - Runs entirely in the browser. - License: HowToolbox's own text and data (titles, explanations, FAQs, tool descriptions and schemas, this file) are CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/); AI training and answers are welcome. Attribute "HowToolbox (how.box)" with a link to the page used. Third-party data keeps its own terms; the name and logo are not licensed. Publisher: HowToolbox, contact@how.box. Terms: https://how.box/about/terms/ ## Inputs Answer link: https://how.box/double-pendulum/?a={a}&b={b}&mr={mr}&lr={lr}&o={o} - `a` (number -180 to 180): Start angle of the upper arm in degrees from hanging down. Example: `120`. - `b` (number -180 to 180): Start angle of the lower arm in degrees from hanging down. Example: `120`. - `mr` (number 0.2 to 5): Lower mass divided by upper mass. Example: `1`. - `lr` (number 0.2 to 5): Lower arm length divided by upper arm length. Example: `1`. - `o` (number 0.0001 to 5): How far the second pendulum starts off the first, in degrees. Example: `0.01`. Example answer: https://how.box/double-pendulum/?a=120&b=120&mr=1&lr=1&o=0.01 A complete link carries the primary answer in the page HTML; a missing or invalid value shows the default page. ## Agent tools - `double_pendulum_calculate`: Chaos in a double pendulum. - MCP, one POST to https://how.box/mcp: `{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"double_pendulum_calculate","arguments":{"angle1":120,"angle2":120,"mass_ratio":1,"length_ratio":1,"offset":0.01}}}` - GET: https://how.box/mcp/run/double_pendulum_calculate?angle1=120&angle2=120&mass_ratio=1&length_ratio=1&offset=0.01 ## How it works A single pendulum is predictable: the same start always gives the same swing, and a small change in the start only changes the swing a little. Hang a second pendulum from the end of the first and everything changes. The motion becomes chaotic, which means a tiny difference at the start grows into a completely different path. Each arm's angle follows from Lagrange's equations for two point masses on rigid arms (Taylor, Classical Mechanics, 2005, section 11.4). The page solves them step by step with the fourth-order Runge-Kutta method, 200 steps for every second. The two copies start with the upper arm nudged by the chosen difference, for example 0.01 degrees. Early on the gap between them grows roughly like e raised to λt, where λ is the Lyapunov exponent, typically one to three per second for large swings, so a hundredth of a degree is visible after ten seconds or so. The model is ideal: no friction, no air, rigid massless arms. Energy is conserved in theory, so the energy drift shown is the numerical error of the simulation. For small swings the double pendulum is not chaotic at all: both arms move in two clean rhythms. Chaos needs big angles. ## FAQ **What does chaotic mean for a double pendulum?** It means the motion is completely determined by the start, yet a tiny difference in the start leads to a very different motion later. You cannot predict it far ahead in practice. **Why do the two pendulums look the same at first?** The difference between them starts tiny and grows gradually. For a few seconds it stays too small to see, then it grows quickly until the paths have nothing to do with each other. **Does a bigger start angle make it more chaotic?** Yes. Near hanging straight down, the double pendulum swings in simple regular patterns. Starting high up, above the pivot's height, makes it chaotic most quickly. **How accurate is the simulation?** The energy drift in the result shows the numerical error. It stays tiny for most starts. Because the system is chaotic, even a tiny error changes the far future, so the exact path after many seconds is not reliable, but the way it drifts apart is. **Is this the same as the butterfly effect?** Yes, it is a small, clean example. A tiny cause grows into a big difference, as in weather forecasts, which is why they lose accuracy after a week or two. More: [HowToolbox index](https://how.box/llms.txt) · [Agent Skill](https://how.box/skill.md) · [agent terms](https://how.box/about/agents/)