##  [Law of Requisite Variety](/law-requisite-variety-0) 

 Definition

A cybernetics principle stating that effective regulation of a system requires a controller whose repertoire of possible responses (variety) is at least as large as the variety of disturbances or states the system may present; without sufficient controller variety, some disturbances cannot be absorbed or countered.

 

 

 

 

 

 





## Principle

Principle

Control succeeds only if the controller can generate responses that map onto the range of perturbations; therefore one must either increase controller variety, reduce environmental variety, or constrain interactions (e.g., via buffering, hierarchy, or rules) to achieve regulation.

 

 

 

 

 





## Demonstration

Demonstration

Illustrative scenario — Situation: A simple on/off regulator designed for two discrete environmental states is applied to an environment with continuous fluctuations. Recognition: The regulator lacks sufficient response variety to stabilize all states. Action: Designer replaces the regulator with a proportional controller or adds buffering to reduce effective disturbance range. Consequence: Regulation succeeds because controller variety or environmental variety is adjusted to restore the required relationship.

 

 

 

 

## Misapplication

Misapplication

Assuming mere addition of components or options always solves variety mismatch (i.e., equating more complexity with more effective control) mistakes raw variety for coordinated, usable variety; the error is neglecting integration, observability, and mapping between disturbances and responses.

 

 

 

 

 





## Consequence

Consequence

Designs guided by the law focus effort on matching controller capabilities to environmental complexity: increase decision alternatives, improve sensing, reduce disturbance scope, or impose structure. Failure to respect the law yields persistent unregulated variability or brittle controllers.

 

 

 

 

## Reversal

Reversal

When disturbances are stochastic and objectives are probabilistic, exact one‑to‑one matching of variety is neither necessary nor practical; stochastic control, robust design, or acceptance of bounded error can provide effective regulation without matching raw variety.

 

 

 

 

 





## Boundary

Boundary

Clearly within: feedback control systems and managerial control problems where disturbances are describable as varieties of states and controller responses are enumerable. Boundary case: adaptive systems where the controller learns over time — requisite variety must be considered in a dynamic rather than static sense. Clearly outside: optimization problems that do not involve regulation against disturbance varieties (pure cost minimization absent external perturbations).

 

 

 

 

 





## Semantic Tension

Semantic Tension

Robustness (ability to absorb disturbances) ↔ Simplicity (minimal, comprehensible controllers): increasing variety can improve robustness but at the cost of complexity, observability and maintainability.

 

 

 

 

 





## Synthesis

Synthesis

Effective regulation requires matching the actionable diversity of responses to the diversity of disturbances; practitioners can either expand coordinated response options, reduce environmental variety by design, or rearrange system architecture (hierarchies, buffers) so that the requisite relationship holds in practice.