Hypotheticals and Counterfactuals
Table of Contents
Workflow
Copy this checklist and track your progress:
Hypotheticals & Counterfactuals Progress:
- [ ] Step 1: Define the focal question
- [ ] Step 2: Generate counterfactuals or scenarios
- [ ] Step 3: Develop each scenario
- [ ] Step 4: Identify implications and insights
- [ ] Step 5: Extract actions or decisions
- [ ] Step 6: Monitor and update
Step 1: Define the focal question
What are you exploring? Past decision (counterfactual)? Future possibility (hypothetical)? Assumption to test? See resources/template.md.
Step 2: Generate counterfactuals or scenarios
Counterfactual: Change one key factor, ask "what would have happened?" Hypothetical: Imagine future scenarios (2-4 plausible alternatives). See resources/template.md and resources/methodology.md.
Step 3: Develop each scenario
Describe what's different, trace implications, identify key assumptions. Make it vivid and concrete. See resources/template.md and resources/methodology.md.
Step 4: Identify implications and insights
What does each scenario teach? What assumptions are tested? What risks revealed? See resources/methodology.md.
Step 5: Extract actions or decisions
What should we do differently based on these scenarios? Hedge against downside? Prepare for upside? See resources/template.md.
Step 6: Monitor and update
Track which scenario is unfolding. Update plans as reality diverges from expectations. See resources/methodology.md.
Validate using resources/evaluators/rubric_hypotheticals_counterfactuals.json. Minimum standard: Average score ≥ 3.5.
Common Patterns
Pattern 1: Pre-Mortem (Prospective Hindsight)
- Format: Imagine it's future date, project failed. List reasons why.
- Best for: Project planning, risk identification before launch
- Process: (1) Set future date, (2) Assume failure, (3) List causes, (4) Prioritize top 3-5 risks, (5) Mitigate now
- When: Before major launch, strategic decision, resource commitment
- Output: Risk list with mitigations
Pattern 2: Counterfactual Causal Analysis
- Format: "What would have happened if we had done X instead of Y?"
- Best for: Learning from past decisions, understanding what mattered
- Process: (1) Identify decision, (2) Imagine alternative, (3) Trace different outcome, (4) Identify causal factor
- When: Post-mortem, retrospective, learning from success/failure
- Output: Causal hypothesis and tests needed; an imagined alternative does not establish causation
Pattern 3: Three Scenarios (Optimistic, Baseline, Pessimistic)
- Format: Describe best case, expected case, worst case futures
- Best for: Strategic planning, forecasting, resource allocation
- Process: (1) Define time horizon, (2) Describe three futures, (3) Assign probabilities, (4) Plan for each
- When: Annual planning, market uncertainty, investment decisions
- Output: Three detailed scenarios with implications
Pattern 4: 2×2 Scenario Matrix
- Format: Two key uncertainties create four quadrants (scenarios)
- Best for: Exploring interaction of two critical unknowns
- Process: (1) Identify two key uncertainties, (2) Define extremes, (3) Develop four scenarios, (4) Name each world
- When: Strategic planning with multiple drivers of uncertainty
- Output: Four distinct future worlds with narratives
Pattern 5: Assumption Reversal
- Format: "What if our key assumption is backwards?"
- Best for: Challenging mental models, unlocking innovation
- Process: (1) List key assumptions, (2) Reverse each, (3) Explore implications, (4) Identify if reversal plausible
- When: Stuck in conventional thinking, need breakthrough
- Output: New perspectives, potential pivots
Pattern 6: Stress Test (Extreme Scenarios)
- Format: Push key variables to extremes, test if decision holds
- Best for: Risk management, decision robustness testing
- Process: (1) Identify decision, (2) List key variables, (3) Set to extremes, (4) Check if decision still valid
- When: High-stakes decisions, need to ensure resilience
- Output: Decision validation or hedges needed
Guardrails
Key requirements:
Plausibility constraint: Scenarios must be possible, not just imaginable. "What if gravity reversed?" is not useful counterfactual. Stay within bounds of plausibility given current knowledge.
Minimal rewrite principle (counterfactuals): Change as little as possible. "What if we had chosen Y instead of X?" not "What if we had chosen Y and market doubled and competitor failed?" Isolate causal factor.
Avoid hindsight bias: Pre-mortem assumes failure, but don't just list things that went wrong in similar past failures. Generate new failure modes specific to this context.
Specify mechanism: Don't just state outcome ("sales would be higher"), explain HOW ("sales would be higher because lower price → higher conversion → more customers despite lower margin").
Probability discipline: Assign probabilities only to clearly defined mutually exclusive, collectively exhaustive outcomes with a defensible basis. Exploratory scenario narratives may remain unweighted; neither equal weights nor forced unequal weights are evidence.
Time horizon clarity: Specify WHEN in future. "Product fails" is vague. "In 6 months, adoption <1000 users" is concrete. Enables tracking.
Extract actions, not just stories: Scenarios are useless without implications. Always end with "so what should we do?" Prepare, hedge, pivot, or double-down.
Update scenarios: Reality evolves. Quarterly review: which scenario is unfolding? Update probabilities and plans accordingly.
Common pitfalls:
- ❌ Confusing counterfactual with fantasy: "What if we had $100M funding from start?" vs. realistic "What if we had raised $2M seed instead of $1M?"
- ❌ Too many scenarios: 10 scenarios = analysis paralysis. Stick to 2-4 meaningful, distinct futures.
- ❌ Scenarios too similar: Three scenarios that differ only in magnitude (10% growth, 15% growth, 20% growth). Need qualitatively different worlds.
- ❌ No causal mechanism: "Sales would be 2× higher" without explaining why. Must specify how change leads to outcome.
- ❌ Hindsight bias in pre-mortem: Just listing past failures. Need to imagine new, context-specific risks.
- ❌ Ignoring low-probability, high-impact: "Black swan won't happen" until it does. Include tail risks.
Quick Reference
Counterfactual vs. Hypothetical:
| Type |
Direction |
Question |
Purpose |
Example |
| Counterfactual |
Backward (past) |
"What would have happened if...?" |
Understand causality, learn from past |
"What if we had launched in EU first?" |
| Hypothetical |
Forward (future) |
"What could happen if...?" |
Explore futures, prepare for uncertainty |
"What if competitor launches free tier?" |
Scenario types:
| Type |
# Scenarios |
Structure |
Best For |
| Three scenarios |
3 |
Optimistic, Baseline, Pessimistic |
General forecasting, strategic planning |
| 2×2 matrix |
4 |
Two uncertainties create quadrants |
Exploring interaction of two drivers |
| Cone of uncertainty |
Continuous |
Range widens over time |
Long-term planning (5-10 years) |
| Pre-mortem |
1 |
Imagine failure, list causes |
Risk identification before launch |
| Stress test |
2-4 |
Extreme scenarios (best/worst) |
Decision robustness testing |
Pre-mortem process (6 steps):
- Set future date: "It's 6 months from now..."
- Assume failure: "...the project has failed completely."
- Individual brainstorm: Each person writes 3-5 reasons (5 min, silent)
- Share and consolidate: Round-robin sharing, group similar items
- Vote on top risks: Dot voting or force-rank top 5 causes
- Mitigate now: For each top risk, assign owner and mitigation action
2×2 Scenario Matrix (example):
Uncertainties: (1) Market adoption rate, (2) Regulatory environment
|
Slow Adoption |
Fast Adoption |
| Strict Regulation |
"Constrained Growth" |
"Regulated Scale" |
| Loose Regulation |
"Patient Build" |
"Wild West Growth" |
Assumption reversal questions:
- "What if our biggest advantage is actually a liability?"
- "What if the problem we're solving isn't the real problem?"
- "What if our target customer is wrong?"
- "What if cheaper/slower is better than premium/fast?"
- "What if we're too early/too late, not right on time?"
Inputs required:
- Focal decision or event: What are you analyzing?
- Key uncertainties: What factors most shape outcomes?
- Time horizon: How far into future/past?
- Constraints: What must remain fixed vs. what can vary?
- Stakeholders: Who should contribute scenarios?
Outputs produced:
counterfactual-analysis.md: Alternative history analysis with causal insights
pre-mortem-risks.md: List of potential failure modes and mitigations
scenarios.md: 2-4 future scenarios with narratives and implications
action-plan.md: Decisions and preparations based on scenario insights
1---2name: what-if-scenario-analysis3description: >-4 Runs what-if analysis on a decision or past event: counterfactuals ("what if we had..."), three-5 scenario and 2x2 scenario sets, assumption reversal and stress tests against alternative6 futures. Use when the user asks "what would have happened if", "what if X changes", wants7 scenarios for a decision, to test assumptions or a thought experiment. Not for a premortem on a8 probability forecast (use forecast-premortem).9metadata:10 source: https://skillmd.com/skills/the-utopia-studio/what-if-scenario-analysis11---1213# Hypotheticals and Counterfactuals1415## Table of Contents16- [Workflow](#workflow)17- [Common Patterns](#common-patterns)18- [Guardrails](#guardrails)19- [Quick Reference](#quick-reference)2021## Workflow2223Copy this checklist and track your progress:2425```26Hypotheticals & Counterfactuals Progress:27- [ ] Step 1: Define the focal question28- [ ] Step 2: Generate counterfactuals or scenarios29- [ ] Step 3: Develop each scenario30- [ ] Step 4: Identify implications and insights31- [ ] Step 5: Extract actions or decisions32- [ ] Step 6: Monitor and update33```3435**Step 1: Define the focal question**3637What are you exploring? Past decision (counterfactual)? Future possibility (hypothetical)? Assumption to test? See [resources/template.md](resources/template.md#focal-question-template).3839**Step 2: Generate counterfactuals or scenarios**4041Counterfactual: Change one key factor, ask "what would have happened?" Hypothetical: Imagine future scenarios (2-4 plausible alternatives). See [resources/template.md](resources/template.md#scenario-generation-template) and [resources/methodology.md](resources/methodology.md#1-counterfactual-reasoning).4243**Step 3: Develop each scenario**4445Describe what's different, trace implications, identify key assumptions. Make it vivid and concrete. See [resources/template.md](resources/template.md#scenario-development-template) and [resources/methodology.md](resources/methodology.md#2-scenario-planning-techniques).4647**Step 4: Identify implications and insights**4849What does each scenario teach? What assumptions are tested? What risks revealed? See [resources/methodology.md](resources/methodology.md#3-extracting-insights-from-scenarios).5051**Step 5: Extract actions or decisions**5253What should we do differently based on these scenarios? Hedge against downside? Prepare for upside? See [resources/template.md](resources/template.md#action-extraction-template).5455**Step 6: Monitor and update**5657Track which scenario is unfolding. Update plans as reality diverges from expectations. See [resources/methodology.md](resources/methodology.md#4-monitoring-and-adaptation).5859Validate using [resources/evaluators/rubric_hypotheticals_counterfactuals.json](resources/evaluators/rubric_hypotheticals_counterfactuals.json). **Minimum standard**: Average score ≥ 3.5.6061## Common Patterns6263**Pattern 1: Pre-Mortem (Prospective Hindsight)**64- **Format**: Imagine it's future date, project failed. List reasons why.65- **Best for**: Project planning, risk identification before launch66- **Process**: (1) Set future date, (2) Assume failure, (3) List causes, (4) Prioritize top 3-5 risks, (5) Mitigate now67- **When**: Before major launch, strategic decision, resource commitment68- **Output**: Risk list with mitigations6970**Pattern 2: Counterfactual Causal Analysis**71- **Format**: "What would have happened if we had done X instead of Y?"72- **Best for**: Learning from past decisions, understanding what mattered73- **Process**: (1) Identify decision, (2) Imagine alternative, (3) Trace different outcome, (4) Identify causal factor74- **When**: Post-mortem, retrospective, learning from success/failure75- **Output**: Causal hypothesis and tests needed; an imagined alternative does not establish causation7677**Pattern 3: Three Scenarios (Optimistic, Baseline, Pessimistic)**78- **Format**: Describe best case, expected case, worst case futures79- **Best for**: Strategic planning, forecasting, resource allocation80- **Process**: (1) Define time horizon, (2) Describe three futures, (3) Assign probabilities, (4) Plan for each81- **When**: Annual planning, market uncertainty, investment decisions82- **Output**: Three detailed scenarios with implications8384**Pattern 4: 2×2 Scenario Matrix**85- **Format**: Two key uncertainties create four quadrants (scenarios)86- **Best for**: Exploring interaction of two critical unknowns87- **Process**: (1) Identify two key uncertainties, (2) Define extremes, (3) Develop four scenarios, (4) Name each world88- **When**: Strategic planning with multiple drivers of uncertainty89- **Output**: Four distinct future worlds with narratives9091**Pattern 5: Assumption Reversal**92- **Format**: "What if our key assumption is backwards?"93- **Best for**: Challenging mental models, unlocking innovation94- **Process**: (1) List key assumptions, (2) Reverse each, (3) Explore implications, (4) Identify if reversal plausible95- **When**: Stuck in conventional thinking, need breakthrough96- **Output**: New perspectives, potential pivots9798**Pattern 6: Stress Test (Extreme Scenarios)**99- **Format**: Push key variables to extremes, test if decision holds100- **Best for**: Risk management, decision robustness testing101- **Process**: (1) Identify decision, (2) List key variables, (3) Set to extremes, (4) Check if decision still valid102- **When**: High-stakes decisions, need to ensure resilience103- **Output**: Decision validation or hedges needed104105## Guardrails106107**Key requirements:**1081091. **Plausibility constraint**: Scenarios must be possible, not just imaginable. "What if gravity reversed?" is not useful counterfactual. Stay within bounds of plausibility given current knowledge.1101112. **Minimal rewrite principle** (counterfactuals): Change as little as possible. "What if we had chosen Y instead of X?" not "What if we had chosen Y and market doubled and competitor failed?" Isolate causal factor.1121133. **Avoid hindsight bias**: Pre-mortem assumes failure, but don't just list things that went wrong in similar past failures. Generate new failure modes specific to this context.1141154. **Specify mechanism**: Don't just state outcome ("sales would be higher"), explain HOW ("sales would be higher because lower price → higher conversion → more customers despite lower margin").1161175. **Probability discipline:** Assign probabilities only to clearly defined mutually exclusive, collectively exhaustive outcomes with a defensible basis. Exploratory scenario narratives may remain unweighted; neither equal weights nor forced unequal weights are evidence.1181196. **Time horizon clarity**: Specify WHEN in future. "Product fails" is vague. "In 6 months, adoption <1000 users" is concrete. Enables tracking.1201217. **Extract actions, not just stories**: Scenarios are useless without implications. Always end with "so what should we do?" Prepare, hedge, pivot, or double-down.1221238. **Update scenarios**: Reality evolves. Quarterly review: which scenario is unfolding? Update probabilities and plans accordingly.124125**Common pitfalls:**126127- ❌ **Confusing counterfactual with fantasy**: "What if we had $100M funding from start?" vs. realistic "What if we had raised $2M seed instead of $1M?"128- ❌ **Too many scenarios**: 10 scenarios = analysis paralysis. Stick to 2-4 meaningful, distinct futures.129- ❌ **Scenarios too similar**: Three scenarios that differ only in magnitude (10% growth, 15% growth, 20% growth). Need qualitatively different worlds.130- ❌ **No causal mechanism**: "Sales would be 2× higher" without explaining why. Must specify how change leads to outcome.131- ❌ **Hindsight bias in pre-mortem**: Just listing past failures. Need to imagine new, context-specific risks.132- ❌ **Ignoring low-probability, high-impact**: "Black swan won't happen" until it does. Include tail risks.133134## Quick Reference135136**Counterfactual vs. Hypothetical:**137138| Type | Direction | Question | Purpose | Example |139|------|-----------|----------|---------|---------|140| **Counterfactual** | Backward (past) | "What would have happened if...?" | Understand causality, learn from past | "What if we had launched in EU first?" |141| **Hypothetical** | Forward (future) | "What could happen if...?" | Explore futures, prepare for uncertainty | "What if competitor launches free tier?" |142143**Scenario types:**144145| Type | # Scenarios | Structure | Best For |146|------|-------------|-----------|----------|147| **Three scenarios** | 3 | Optimistic, Baseline, Pessimistic | General forecasting, strategic planning |148| **2×2 matrix** | 4 | Two uncertainties create quadrants | Exploring interaction of two drivers |149| **Cone of uncertainty** | Continuous | Range widens over time | Long-term planning (5-10 years) |150| **Pre-mortem** | 1 | Imagine failure, list causes | Risk identification before launch |151| **Stress test** | 2-4 | Extreme scenarios (best/worst) | Decision robustness testing |152153**Pre-mortem process** (6 steps):1541551. **Set future date**: "It's 6 months from now..."1562. **Assume failure**: "...the project has failed completely."1573. **Individual brainstorm**: Each person writes 3-5 reasons (5 min, silent)1584. **Share and consolidate**: Round-robin sharing, group similar items1595. **Vote on top risks**: Dot voting or force-rank top 5 causes1606. **Mitigate now**: For each top risk, assign owner and mitigation action161162**2×2 Scenario Matrix** (example):163164**Uncertainties**: (1) Market adoption rate, (2) Regulatory environment165166| | Slow Adoption | Fast Adoption |167|---------------------|---------------|---------------|168| **Strict Regulation** | "Constrained Growth" | "Regulated Scale" |169| **Loose Regulation** | "Patient Build" | "Wild West Growth" |170171**Assumption reversal questions:**172173- "What if our biggest advantage is actually a liability?"174- "What if the problem we're solving isn't the real problem?"175- "What if our target customer is wrong?"176- "What if cheaper/slower is better than premium/fast?"177- "What if we're too early/too late, not right on time?"178179**Inputs required:**180- **Focal decision or event**: What are you analyzing?181- **Key uncertainties**: What factors most shape outcomes?182- **Time horizon**: How far into future/past?183- **Constraints**: What must remain fixed vs. what can vary?184- **Stakeholders**: Who should contribute scenarios?185186**Outputs produced:**187- `counterfactual-analysis.md`: Alternative history analysis with causal insights188- `pre-mortem-risks.md`: List of potential failure modes and mitigations189- `scenarios.md`: 2-4 future scenarios with narratives and implications190- `action-plan.md`: Decisions and preparations based on scenario insights