Chaos Lab helps users test how a system behaves when small changes are repeated over time. It is useful for studying stability, feedback, thresholds, and sensitivity. The goal is not just to show mathematics. The goal is to help users reason about real systems that can settle, oscillate, or become unstable.
Chaos Lab simulates a simple nonlinear rule over many iterations and shows what happens over time. It helps users answer questions such as:
Core equation used by the system
You do not solve this once. The system applies the rule repeatedly. Each output becomes the next input.
Useful for teaching deterministic chaos, nonlinear systems, growth limits, and the difference between randomness and rule-driven unpredictability.
Useful for reasoning about instability, feedback loops, sensitivity, and why small timing or state differences can grow into different outcomes.
Useful for studying thresholds, transitions, divergence, and simplified models of bounded growth.
Useful for understanding adoption growth, saturation, load pressure, and systems where increasing feedback can push operations into instability.
The best way to use Chaos Lab is to begin with a use case. Each use case answers a practical question and helps users decide what values to enter.
Use this when you want to know whether a system settles into a steady long-run state.
A fish population in a lake, a resource consumption cycle, or a system load that should level off rather than swing wildly.
If I start from the current state and apply the same growth rule repeatedly, does the system settle?
How to derive x₀: Normalize the current state against the practical maximum.
Example: a lake supports 1000 fish and currently has 400 fish.
How to think about r: Use a lower value when growth is present but naturally constrained. This represents moderate growth pressure with some stabilizing effect.
What to look for: A sequence graph that settles and a negative Lyapunov exponent.
Use this when you want to know whether a system repeats in cycles instead of settling at one level.
Seasonal demand cycles, recurring operational peaks, or systems that bounce between high and low states under repeated feedback.
Does the system fall into a repeating pattern rather than a single stable value?
How to derive x₀: Use the current normalized level of the system.
How to think about r: Use a moderate-to-higher feedback value when the system reacts strongly but still remains bounded.
Example: an operational platform is running at 40 percent of intended peak and tends to surge in recurring patterns.
What to look for: Repeating patterns in the sequence graph and behavior that does not flatten to one value.
Use this when you want to test whether two very similar starting states become very different later.
Forecasting, operational planning, adoption modeling, or any system where small early differences may change long-run outcomes.
If I start with almost the same initial state, will the outcomes remain close or diverge rapidly?
How to derive x₀: Start with the current normalized state.
How to derive comparison x₀: Use a nearly identical value to represent a tiny measurement or timing difference.
How to think about r: Use a higher feedback value when you are testing whether amplified response creates instability.
What to look for: The divergence graph is the primary output here. If the gap grows quickly, the system is sensitive.
Use this when you want to know where the system changes type. This is useful for finding boundaries between stable, periodic, and chaotic behavior.
Capacity planning, growth stress testing, or policy tuning where increasing pressure eventually pushes the system into a less predictable state.
At what control level does the system stop behaving well and enter a different regime?
How to derive x₀: Use the current normalized system level.
How to think about r: Sweep across a range of values to represent increasing pressure, growth, or amplification.
Example: a service currently runs at 45 percent effective load, but you want to know at what control intensity the system becomes unstable.
What to look for: The regime map and classification changes are the most useful outputs.
Use this when you want to explore how strong feedback can destabilize a bounded system.
Product adoption loops, network effects, operational surges, demand feedback, or simplified market-like amplification behavior.
What happens when feedback becomes strong enough to amplify small changes across repeated periods?
How to derive x₀: Normalize the current share, load, or adoption level.
Example: 12,000 out of 30,000 possible users have adopted a product.
How to think about r: Let r represent the strength of amplification. This could mean referral power, demand reaction, or growth acceleration.
What to look for: Divergence growth, unstable sequences, and positive Lyapunov values.
This use case is useful for reasoning. It is not a promise of real-world prediction.
The strongest practical value of this system comes from translating real conditions into model inputs carefully. Use the following method across all use cases.
Choose something that has a current level and a practical maximum. Examples include fish population, active users, safe server load, adoption share, or resource utilization.
Convert the real value into a number between 0 and 1.
Do not think of r as an abstract number only. Treat it as a simplified measure of how strongly the current state affects the next state. Higher r means stronger growth or stronger amplification.
Iterations can represent days, cycles, quarters, updates, or repeated decision periods. Use more iterations when you want to see long-run behavior.
Use this to see whether the system settles, repeats, or becomes irregular over time.
Use this to see whether small starting differences remain small or grow into large gaps.
Use this as the main quantitative clue. Negative suggests stability. Positive suggests chaos-like divergence.
Use this when you want to see where a small parameter change pushes the system into a new behavior class.
Start with the use case that best matches your problem. Use the presets first, then adjust values based on your real-world scenario and observe how the system responds.
Launch Chaos Lab