---
slug: "WeatherAgent"
name: "WeatherAgent"
type: "Strategy"
layer: "Mind"
ref: "WeatherAgent#25e1"
sema_id: "sema:WeatherAgent#mh:SHA-256:25e1d749c629abfedc88613c936da6d6aa40e661d411d6764fe70b30ae690274"
tags: ["Mind","Strategy"]
updated: 2026-11-11
mindscape: "origin-of-risk"
index_url: https://mindscape.pages.dev/origin-of-risk/index.json
url: https://mindscape.pages.dev/origin-of-risk/WeatherAgent/
related: ["FractalityTest","InformationDividend","ShapeControl","VectorOptionTimeline"]
---

# WeatherAgent

Choose between holding shape (adding inertia) and moving with the information (removing inertia) by weighing switching cost, information reliability and local variation

Layer: Mind · Category: Strategy

## Mechanism

An agent can be in one of several places (Home or Park) and the payoff of each depends on the state of the world (Rain or Sun); moving between them costs something. The same setup can be seen two ways at once, like a gestalt figure: as [ShapeControl](https://mindscape.pages.dev/origin-of-risk/ShapeControl.md) (a self-righting agent keeping its payoff maximal whatever the weather) or as exploiting information by changing shape. The decision rule follows from three variables. If short-term information is unreliable, add inertia: move less and wait for more evidence. If it is reliable, switching costs are low and the [FractalityTest](https://mindscape.pages.dev/origin-of-risk/FractalityTest.md) shows local conditions differ from the average, remove inertia and move often to collect the [InformationDividend](https://mindscape.pages.dev/origin-of-risk/InformationDividend.md). If switching costs are high, hold one shape chosen from the long-range forecast. Lowering the cost of moving is itself a strategy for raising the dividend.

## Invariants

- Both views, maintaining shape and exploiting information by changing it, apply to the same agent
- More inertia is right for noisy information; less inertia for reliable, locally varying information
- The dividend from moving is capped by the cost of moving

## Preconditions

- The agent can change position or policy at a known cost
- Some short-range information about the current state is available

## Failure modes

- Treating instability as noise when it is reliable information that averages out over the long range
- Moving on every signal when signals are unreliable
- Not investing in lowering switching costs when the dividend would justify it

## Builds on

- [FractalityTest](https://mindscape.pages.dev/origin-of-risk/FractalityTest.md) (Mind; References)
- [InformationDividend](https://mindscape.pages.dev/origin-of-risk/InformationDividend.md) (Mind; References)
- [ShapeControl](https://mindscape.pages.dev/origin-of-risk/ShapeControl.md) (Mind; References)

## Used by

- [VectorOptionTimeline](https://mindscape.pages.dev/origin-of-risk/VectorOptionTimeline.md) (Society; References)

## Identity

- Sema ID: `sema:WeatherAgent#mh:SHA-256:25e1d749c629abfedc88613c936da6d6aa40e661d411d6764fe70b30ae690274`
- Ref: `WeatherAgent#25e1`

## Links

- [Pattern page](https://mindscape.pages.dev/origin-of-risk/WeatherAgent/)
- [View in the interactive graph](https://mindscape.pages.dev/origin-of-risk/#WeatherAgent)
- [JSON](https://mindscape.pages.dev/origin-of-risk/WeatherAgent.json)
- [Index of all patterns](https://mindscape.pages.dev/origin-of-risk/index.json)