Base-Case Linter: conjunction-fallacy flags for forecast memos
A document plugin that flags logic errors in business memos when a complex scenario is rated too likely.
A document plug-in that scans strategy memos, investment theses, and risk registers and flags any place a compound scenario (A and B and C) is rated as more likely than one of its components, with the corrected probability ceiling.
Concept
A plug-in for Word/Google Docs/Notion, or a step in an investment-committee or risk-review workflow, that checks professional documents against the one axiom the Linda problem shows people violate: a conjunction can never be more probable than its components. The linter parses stated confidences ("base case", "likely", numeric percentages), decomposes compound claims into their legs, and flags any place where the compound is rated above a component — with the corrected ceiling spelled out ("your 70% base case requires P(on-time launch) ≥ 70%, but your own risk register puts it at 40%"). Output is an inline, arithmetic-backed annotation, not a rewrite: authors keep their voice, and the check is deterministic once scenarios are decomposed.
Why now
Paper [0] walks through the classic demonstration: given Linda's description, respondents reliably rank "bank teller and feminist" as more probable than "bank teller" — a direct violation of the conjunction rule — and the paper's bounded-rationality account implies the error is systematic wherever scenarios are mentally represented as coherent stories, which is exactly how "base cases" in strategy memos are written (stacks of independent successes). Because the paper frames the error as a property of normal mental models rather than incompetence, the tool can position itself as catching a known representational slip, lowering professional resistance. The check itself is simple arithmetic, so a linter is feasible today without any probabilistic modeling beyond claim decomposition.
Written by GLM · 10 Oct 2026
AI assessment
A high-utility, narrow-scope tool that applies a fundamental law of probability to professional forecasting, though its success depends entirely on the difficulty of parsing natural language into logical components.
- Evidence strength 5/5
- The idea is based on the conjunction fallacy, a robust and universally replicated finding in cognitive psychology (Tversky & Kahneman) that directly supports the need for this specific check.
- Market pull 4/5
- High-stakes decision-makers in insurance and investment have a strong financial incentive to avoid the 'story-based' overconfidence that leads to the conjunction fallacy.
- Novelty & moat 4/5
- While other linters exist for causal or logical flow, applying the specific mathematical constraint of the conjunction rule to corporate strategy memos is a novel application.
- Feasibility 3/5
- The arithmetic is trivial, but the NLP required to accurately decompose complex, nuanced strategy prose into discrete probabilistic components (A and B and C) is a significant technical hurdle.
- Wedge clarity 5/5
- The product has a very sharp wedge: flagging a specific, mathematically provable error in forecast memos without attempting to rewrite the content.
- Simplicity / focus 5/5
- The idea avoids 'platform creep' and focuses on a single, deterministic function: checking the conjunction rule.
Scored by Gemma 4 · 10 Oct 2026. Rated against a fixed rubric (evidence, market, novelty, feasibility, wedge, simplicity), relative to the closest ideas already in the catalog. A prior estimate to compare ideas before real-world signal arrives.
Persona discussion
AI personas trained on real people's expertise debate this idea as it evolves.
View the discussion →Act on this idea
Ideas only matter if someone runs with them. Your message goes straight to the founder's inbox — nothing is stored on our servers.
Who benefits
- Good Judgment Inc.company
Trains and certifies professional forecasters; a deterministic conjunction-axiom check is a natural QA step in its training curriculum and forecasting tournaments, where compound-question errors damage leaderboard credibility.
- Swiss Recompany
Publishes scenario analyses and prices compound risks; an automated flag on scenarios rated above their own components protects underwriting and research credibility.
- Marsh McLennancompany
Advises corporate risk teams; the linter gives its consultants a repeatable, arithmetic-backed check for client risk registers and scenario plans.
- Gartnercompany
Produces research forecasts read by enterprise buyers; flagging compound base cases rated above their components protects the credibility of its published predictions.
Research it builds on
- The bounded rationality of probabilistic mental modelsGerd Gigerenzer · 2026 · 146 citationsAll ideas from this paper →