# Analyze Counterfactual Decisions

Teams often struggle to learn from decisions because the unchosen path stays vague. This skill creates structured counterfactual comparisons with scenarios, probabilities, and practical lessons.

## Install

```bash
npx skillstore add bellabe/reasoning-counterfactual
```

## Metadata

- - Slug: bellabe-reasoning-counterfactual
- - Version: 1.0.0
- - Author: BellaBe
- - GitHub username: BellaBe
- - License: MIT
- - Repository: https://github.com/BellaBe/lean-os/tree/main/.claude/skills/reasoning-counterfactual
- - Ref: main
- - Supported tools: Claude, Codex, Claude Code
- - Risk level: safe
- - Quality score: 80
- - Quality tier: silver
- - Public page: https://skillstore.pages.dev/skills/bellabe-reasoning-counterfactual
- - Manifest: https://skillstore.pages.dev/api/skills/bellabe-reasoning-counterfactual/manifest

## Capabilities

- Documents the actual decision, observed outcome, metrics, and causal chain.
- Defines a specific intervention that changes one decision variable.
- Projects expected, optimistic, and pessimistic scenarios with probabilities.
- Compares actual and alternative outcomes using quantitative and qualitative factors.
- Extracts lessons, recommendations, and next actions from the comparison.

## Use Cases

- Review a strategic decision: Compare a past strategy with a realistic alternative and identify what should change next time.
- Plan a future scenario: Model likely outcomes if a market, competitor, or internal choice changes.
- Improve postmortems: Separate hindsight bias from actionable lessons after a project, launch, or investment decision.

## Prompt Templates

### Analyze one past decision

```
Use counterfactual reasoning to evaluate this decision: [decision]. Actual outcome: [outcome]. Alternative to test: [alternative].
```

### Compare two paths

```
Compare our actual path with this alternative path. Include metrics, causal assumptions, weighted scenarios, and a practical recommendation.
```

### Evaluate a strategy shift

```
Analyze whether we should have chosen [strategy A] instead of [strategy B]. Use expected, optimistic, and pessimistic cases with probabilities.
```

### Run a decision retrospective

```
Create a full counterfactual retrospective for this decision. Include the actual world, intervention validity, alternative projection, comparison, insight, and next actions.
```

## Limitations

- It depends on accurate facts about the original decision and outcome.
- It cannot prove what would have happened with certainty.
- It is less useful when the alternative was never feasible.
- It does not run experiments or collect external data by itself.

## Best Practices

- Use facts that were knowable at the time of the decision.
- Change one main variable before adding complex multi-variable scenarios.
- Assign probabilities and state uncertainty for every projected outcome.

## Anti Patterns

- Do not treat projections as proof of what would have happened.
- Do not test alternatives that were impossible or unavailable.
- Do not ignore negative downstream effects of the preferred option.

## Security Audit

- - Safe to publish: true
- - Audited at: 2026-06-28T15:33:56.954\+00:00
- - Summary: Static analysis flagged Markdown code fences and ordinary words as shell execution, weak cryptography, and system reconnaissance. Manual review found a single instructional Markdown file with YAML examples, decision-analysis guidance, and no executable code or prompt-injection attempt.

## Stats

- - Views: 221
- - Downloads: 4
- - Favorites: 0
- - Popularity score: 0