What is nonlinearity?

Nonlinearity means that doubling a cause does not double the effect. Many relationships in economies and systems bend, saturate, or jump. Treating them as straight lines leads to forecasts that are fine near the starting point and badly wrong away from it.

How it works

Nonlinear effects arise from thresholds, saturation, diminishing returns, and feedback loops that change strength as conditions change. A server that handles load smoothly up to 70 percent of capacity may slow sharply beyond 90 percent. A price cut may raise sales a lot at first and little after that. Nonlinearity is also why combinations matter: two small changes together can do more than the sum of each alone. For modelers, it means testing across the full range of inputs, not only at the current values.

Nonlinearity also changes how to read a forecast. A model built on a straight-line fit may work in normal conditions and fail at the edges, which are exactly the conditions where good decisions matter most.

Example

Say a call center answers calls with a wait time that depends on how busy the agents are. When each call takes 1 minute to handle, the average wait at 50 percent utilization is about 1 minute. At 80 percent it is about 4 minutes, and at 95 percent it is about 19 minutes. Utilization rose by less than a factor of two, but waiting rose by a factor of nearly 20. This shape is typical of queues. This is an illustration with round numbers.

How to model it in Machinations

Use a Register with a formula that bends, such as a power, a cap, or a lookup by ranges, to set a rate. Connect it with a state connection to the flow it controls. Play the diagram while changing the input Lever across a wide range, and chart the output to see where the curve steepens.