Systems Thinking

Use this contract for the concrete task. Apply a short relevant check for a small change; expand investigation when the failure, risk or requested artifact warrants it. Resolve facts from the repository and runtime before imposing a process. Delegation is optional and should answer a bounded unresolved question.

Analysis contract

Define the decision, system boundary, time horizon and observed behavior. Treat feedback/archetype explanations as hypotheses. A recurring bug may still have a direct local cause; do not force an organizational or archetypal explanation.

• Separate observed links, plausible mechanisms and untested assumptions. Look for alternative structures that could generate the same behavior.
• Distinguish stocks, flows and information links; preserve units and time scales. Check polarity of each causal link under stated conditions.
• Record delays, capacity limits, nonlinearities and exogenous influences. Feedback diagrams communicate hypotheses; they do not identify causality by themselves.
• Quantify only when parameters have evidence or defensible bounds. Use sensitivity/scenario ranges when calibration is weak; do not invent precise crisis dates.
• Compare feasible interventions including a no-change case and possible side effects. A high position in a leverage-point hierarchy is not evidence that an intervention is feasible or beneficial.
• Validate trajectories against available observations and revise the model when counterevidence appears. State which predictions would falsify the current explanation.

Evidence and output

Produce the smallest model that informs the decision: causal map, stock-flow equations or scenario comparison. Include boundary, evidence sources, assumptions/uncertainty, alternative explanation and an observation/experiment that discriminates. Use readable plots with honest scales; label qualitative trajectories and illustrative numbers.

Fault-specific references

• Recurring pattern with unclear mechanism: recognizing-system-patterns.md; archetypes are candidate explanations, not diagnoses.
• Feedback link/polarity uncertainty: causal-loop-diagramming.md.
• Accumulation, delay or equilibrium: stocks-and-flows-modeling.md.
• Intervention comparison: leverage-points-mastery.md.
• Communicating a modeled/measured trajectory: behavior-over-time-graphs.md.

Software structure, statistical inference and simulation integration have separate specialist methods. Use them only when the concrete task crosses those boundaries. No fixed learning path, duration or full suite of diagrams is required.

Optional references

All sheets below are in this directory. Choose a sheet because its checks or examples help the task; there is no requirement to read the catalog in sequence. Verify time-sensitive APIs and numerical/performance claims before relying on examples.

• behavior over time graphs
• causal loop diagramming
• leverage points mastery
• recognizing system patterns
• stocks and flows modeling
• systems archetypes reference