reasoning-counterfactual

Counterfactual reasoning lets you model past decisions with hypothetical interventions to compare actual outcomes against alternatives. Structure your scenario using the four-stage process: define the actual world, specify the intervention, project alternative outcomes, then analyze differences with probability weighting to see how changing one decision affects results.

Scenario planning explores multiple futures; counterfactual analysis isolates one specific intervention on a past decision to measure its impact. This Skill uses probability-weighted comparative analysis to show quantified differences between what happened and what could have happened, making hindsight actionable for future decisions.

Yes. You can model interventions like price adjustments, feature changes, or timing shifts on past decisions, then compare projected outcomes such as ARR or customer acquisition against actual results using probability weighting to account for uncertainty in alternative scenarios.

The Skill uses a defined YAML schema to capture actual outcomes, specify interventions, project scenarios, and perform difference analysis. You provide the baseline decision state, describe the intervention precisely, define alternative projections, and the Skill applies minimal intervention constraints and causal tracing to generate comparative insights.

Probability weighting assigns confidence levels to each alternative scenario you model, ensuring the comparative analysis accounts for uncertainty rather than treating all paths as equally likely. This produces more realistic difference analysis when outcomes depend on variables you cannot control.

Counterfactual analysis works for past decisions you cannot re-run experimentally, decisions with long feedback loops, and scenario planning where A/B testing is infeasible. It trades experimental rigor for the ability to extract insights from historical data through structured intervention simulation and causal reasoning.