Skip to content

Token Waste Reinforcing Loop

Run the Token Waste Reinforcing Loop MicroSim Fullscreen

About This MicroSim

A causal loop diagram shows how variables influence one another around closed paths of cause and effect. This one uses the book's own token-waste example from Chapter 9 to put the two basic loop types side by side.

Loop R, Token Waste (reinforcing). More unnecessary parallel agents pay more startup overhead. More overhead leaves fewer tokens remaining in the usage window. Fewer remaining tokens raise the pressure to rush, and rushing launches even more unnecessary agents. The change comes back amplified.

Loop B, Chapter Token Budgeting (balancing). The same overhead, once it pushes a chapter past its allowance, draws more budgeting attention. Budgeting raises cost awareness, and cost awareness reduces unnecessary parallel agents. The change comes back reversed.

The two loops share one link, from Unnecessary Parallel Agents to Startup Overhead Paid. That shared link is where the two loops work against each other.

Link Sign Loop
Unnecessary Parallel Agents → Startup Overhead Paid + R and B
Startup Overhead Paid → Tokens Remaining in Window − R
Tokens Remaining in Window → Pressure to Rush − R
Pressure to Rush → Unnecessary Parallel Agents + R
Startup Overhead Paid → Chapter Token Budgeting + B
Chapter Token Budgeting → Cost Awareness + B
Cost Awareness → Unnecessary Parallel Agents − B

A + means the two variables move in the same direction. A − means they move in opposite directions. Loop R has two minus signs (an even number), so it is reinforcing. Loop B has one (an odd number), so it is balancing.

A model, not a measurement

The diagram is a qualitative model of cause and effect. The only number in it, roughly 12,000 tokens of startup overhead per agent, is the figure measured on this project and reported in Chapter 8.

How to Use

  1. Select R loop, or click the red R circle. The reinforcing loop lights up and each variable shows whether it rises or falls when Unnecessary Parallel Agents rises.
  2. Read the traced change in the panel from left to right. Notice that it returns to the starting variable pointing the same way it started.
  3. Select B loop, or click the blue B circle. Trace the same starting change and notice that this time it returns pointing the opposite way.
  4. Count the minus signs on each loop and check the count against the "How to tell" line.
  5. Hover any variable to read its definition, or any arrow to read the causal claim it makes.
  6. Drag the background to pan. Use the −, +, and Fit buttons to zoom. Pinch also zooms. A normal scroll is left to the page.
  7. Select Show Both to return to the full diagram.

Iframe Embed Code

You can add this MicroSim to any web page by adding this to your HTML:

1
2
3
4
<iframe src="https://dmccreary.github.io/ibook-skills/sims/token-waste-reinforcing-loop/main.html"
        height="562px"
        width="100%"
        scrolling="no"></iframe>

Lesson Plan

Audience

Adult learners, educators, and instructional designers who build intelligent textbooks with AI agent skills. No prior systems-thinking background is needed.

Learning Objective

Differentiate a reinforcing loop from a balancing loop using this book's own token-waste example from Chapter 9. (Bloom's Taxonomy: Analyze)

Duration

15 minutes

Prerequisites

  • Knows that each additional agent pays a fixed startup overhead (Chapter 8)
  • Has read the token waste antipatterns and chapter token budgeting sections of Chapter 9

Activities

  1. Exploration (4 min): Hover all six variables and all seven arrows. For each arrow, say the causal claim in a full sentence that starts with "The more..." or "The fewer...".
  2. Guided Practice (6 min): Before selecting either loop, predict which one is reinforcing by counting minus signs. Then select R loop and B loop and compare the traced change with your prediction.
  3. Assessment (5 min): Answer the questions below in writing.

Assessment

  • The learner states that a reinforcing loop amplifies a change and a balancing loop counteracts it.
  • The learner classifies each loop correctly and justifies the answer with the even or odd count of minus signs.
  • The learner identifies the shared link and explains that it carries the effect of both loops.

Discussion Questions

  1. If the link from Startup Overhead Paid to Chapter Token Budgeting were removed, no author would ever notice an overrun. What would be left, and how would the system behave?
  2. Which single link in loop R would you weaken first, and what practice from Chapter 9 weakens it?
  3. Name another pair of loops, from any field, with the same shape: a runaway habit and the check that restrains it. A pattern that recurs across domains like this is a systems archetype.

References

  1. Causal loop diagram - Wikipedia - Explains link polarity, reinforcing and balancing loops, and the rule of counting negative links.
  2. System archetype - Wikipedia - Describes the recurring patterns of interacting loops that appear across many domains.
  3. Thinking in Systems: A Primer - Donella H. Meadows - A readable introduction to stocks, flows, and feedback loops.
  4. vis-network Documentation - vis.js - Reference for the network library used to draw this diagram.