Riskiest Assumption Test¶
Rank a plan's assumptions by consequence-if-false times uncertainty times upstream-position, then spend the next effort on the cheapest credible test of the top one — the question whose answer would most reduce total wasted work.
Core Idea¶
The riskiest assumption test (RAT), from lean-startup methodology, is the discipline of finding — among all the assumptions a plan depends on — the one combining the largest downside if false with the weakest current evidence, and then designing the cheapest credible experiment to test it before committing downstream resources. The calculation is a product of three terms: consequence if false, probability of being wrong, and upstream position in the dependency graph. The next unit of effort goes to the top of that ranking.
Scope of Application¶
The ordering it codifies — invest the next effort where the result would most change downstream decisions — recurs under its own name across planning fields.
- Lean-startup product development — the named home: a presales page or concierge test before scaling.
- Pharmaceutical R&D — the "killer experiment" run early to avoid a Phase III failure.
- Military planning — red-teaming and pre-mortem rehearsal of the assumed adversary response.
- Aerospace engineering — critical-path testing under the technology-readiness framework.
- Experimental science — Platt's strong inference discriminating between consequential hypotheses.
- Investment due diligence — aiming the diligence budget at the thesis-reversing assumption.
Clarity¶
The name makes legible a distinction planning otherwise hides: testing for evidence versus testing for comfort. It names the convenience trap — running the test the team knows how to run rather than the one whose answer would most change the plan. It also targets the uncertainty endpoint that ordinary risk vocabulary cannot see: not the largest exposure under known statistics, but the least-characterized and most consequential assumption.
Manages Complexity¶
A long, unranked list of assumptions and the overwhelming injunction to "de-risk everything" collapse to a single ordering question. Every assumption is scored on the same three terms, and the next experiment is read off the top cell. The three terms resolve cases "test early" cannot tell apart: consequential-but-supported ranks low, poorly-evidenced-but-mild ranks low, and consequential-but-downstream is deferred until upstream tests resolve.
Abstract Reasoning¶
RAT licenses a diagnostic that ranks assumptions on three terms and reads the next experiment off the top cell, a failure-mode diagnostic that names the convenience trap, an interventionist move that runs the cheapest credible falsification now and predicts the cost of deferral, and boundary-drawing that separates risk from uncertainty, truth-slippage from time-slippage, and the meta-procedure from the instruments and statistics it deploys.
Knowledge Transfer¶
Within product practice the discipline transfers as mechanism with the convenience-trap warning intact, MVP and presales-pricing being the home instruments. Remarkably, the ranking genuinely travels across substrates — but the cross-field lesson belongs to the parent Value of Information (composed with Critical path method, pre_mortem, and an assumption inventory), which pharma, defense, and science already instantiate under their own names. Outside its home "RAT" gets borrowed loosely, flattened to "do early tests" or "MVP everything," which loses the very ordering discipline that distinguishes it.
Relationships to Other Abstractions¶
Current abstraction Riskiest Assumption Test Domain-specific
Parents (3) — outgoing typed relations
-
Riskiest Assumption Test includes Assumption Prime
The objects ranked and tested are propositions the plan treats as true and on which downstream work depends.
-
Riskiest Assumption Test includes Prioritization Prime
RAT ranks competing assumptions by a rule so scarce testing effort goes to the highest-value question first.
-
Riskiest Assumption Test is a decomposition of Value of Information Prime
Removing lean-startup vocabulary leaves the rule of buying the evidence expected to improve downstream decisions most relative to its cost.
Hierarchy paths (12) — routes to 10 parentless roots
- Riskiest Assumption Test → includes Assumption → is a kind of Epistemic Mode Of A Proposition
- Riskiest Assumption Test → includes Prioritization → is a kind of Optimization
- Riskiest Assumption Test → includes Prioritization → is a kind of Preference
- Riskiest Assumption Test → is a decomposition of Value of Information → presupposes Uncertainty
- Riskiest Assumption Test → is a decomposition of Value of Information → presupposes Decision → presupposes Constraint
- Riskiest Assumption Test → is a decomposition of Value of Information → presupposes Decision → presupposes Reversibility and Irreversibility
- Riskiest Assumption Test → includes Prioritization → presupposes Allocation → presupposes Scarcity → is a decomposition of Constraint
- Riskiest Assumption Test → is a decomposition of Value of Information → presupposes Decision → is a decomposition of Stage Gate Process → is a kind of, typical Sequencing → presupposes Dependency
- Riskiest Assumption Test → is a decomposition of Value of Information → includes Expected Value → is a kind of Aggregation → is a kind of Transformation → is a kind of, typical Function (Mapping)
- Riskiest Assumption Test → is a decomposition of Value of Information → presupposes Decision → is a decomposition of Stage Gate Process → is a kind of, typical Sequencing → is a kind of Optimization
- Riskiest Assumption Test → is a decomposition of Value of Information → includes Expected Value → presupposes Probability → is a kind of Measure → presupposes Set and Membership
- Riskiest Assumption Test → is a decomposition of Value of Information → presupposes Decision → is a decomposition of Stage Gate Process → is a kind of, typical Sequencing → presupposes Time
Neighborhood in Abstraction Space¶
Riskiest Assumption Test sits in a sparse region of the domain-specific corpus (75th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Organizational & Operational Failure Modes (38 abstractions)
Nearest neighbors
- Effectuation — 0.83
- Innovation Accounting — 0.83
- Black Elephant — 0.82
- Folk Theorem (Repeated Games) — 0.82
- Parkinson's Law of Triviality (Bikeshedding) — 0.82
Computed from structural-signature embeddings · 2026-10-25