Conditioned Probability Frame Specification

State what is being taken as given before interpreting, comparing, or acting on a probability.

Version

v1 · 2026-08-24 · History

Solution archetype #

208

Problem family

Uncertainty, Evidence & Inference Failure

Problem subfamily

Probability, Distribution & Risk Calibration

Summary

A probability claim is incomplete until it says what is being taken as given. Conditioned Probability Frame Specification makes the hidden context explicit: the event whose probability is being assessed, the evidence or scenario already assumed, the denominator population left after that assumption, and the boundary within which the number may be compared or reused.

The target prime is conditional_probability: re-normalizing a probability measure to the information context taken as given. The archetype is not a generic Bayesian update, a risk score, or a causal-effect estimator. It is the reusable design pattern that makes those downstream operations safe: first declare the conditioning frame, then compute and communicate the probability inside that frame.

When This Archetype Applies

Partial catalog groundingSome structural conditions are represented by existing abstractions, but no sufficient condition set is fully represented.

Diagnostic problem

Probability claims are made, compared, or used in decisions without declaring what information, subgroup, scenario, evidence state, or assumption set is being taken as given. The missing conditioning frame lets actors mix marginal and conditional probabilities, compare rates with different denominators, ignore base rates, reuse numbers outside their valid context, or treat an updated probability as if it were unconditional. The arithmetic may be correct inside one frame while the conclusion is wrong because the frame has silently changed.

Applicability expression6 distinct conditions

Probability estimateandEstimate conditioned on contextandConditional-probability confusionandCross-context probability comparisonandContext-defined denominatorandUnstated evidence context

Algebraic1∧2∧3∧4∧5∧6

groundedpartly groundedopen

6 conditions, all required.

6Required in every casenumbered 1–6

These hold no matter which pattern applies.

1

Probability estimate · grounded · any one of 2

A probability, risk, prevalence, rate, likelihood, forecast, or uncertainty estimate is stated or compared. more

a

primeProbability— Quantifies uncertainty and likelihoods.

or — any one of these establishes this case

b

primeConditional Probability— Re-normalize a probability measure to the information context that is taken as given.

and — also required

2

Estimate conditioned on context · grounded

The estimate depends on observed evidence, subgroup membership, a scenario, signal, sampling process, or selection rule. more

primeConditional Probability— Re-normalize a probability measure to the information context that is taken as given.

and — also required

3

Conditional-probability confusion · open

Decision-makers may confuse marginal, joint, posterior, forward-conditional, and inverse-conditional probabilities. more

and — also required

4

Cross-context probability comparison · open

Probability claims are compared across groups, models, periods, jurisdictions, experiments, or data slices. more

and — also required

5

Context-defined denominator · open

The relevant denominator is a context-defined subset of the visible population. more

and — also required

6

Unstated evidence context · open

A decision depends on whether evidence should update a prior, but the evidence context is unstated. more

Other requirements and context (2)

Why these sit outside the expression

Supporting context — it may accompany or help interpret the situation, but it is not a load-bearing condition in a sufficient diagnostic set.

• Supporting contextA probability will be reused downstream by people who may not know the reference population, information state, or assumptions used to produce it. more

• Supporting contextThe cost of a denominator error is high enough that the probability statement needs auditability. more

Coverage

2 of 6 conditions grounded · 4 open.

Read the methodology · Download the trigger-logic data

Structural problem

A naked probability number invites denominator mistakes. “Ten percent risk,” “ninety percent accuracy,” or “high likelihood” can mean very different things depending on the population, evidence, scenario, time window, and sampling process. When those conditions are hidden, people compare marginal and conditional rates, invert conditional probabilities, ignore base rates, or move estimates across contexts where they no longer apply.

Core intervention

The intervention is to attach a probability claim to a complete conditioning frame. The frame includes a target event, a stated given-context, a denominator population or probability mass, a renormalization rule, a probability-type label, a compatibility check for comparison, and a reuse boundary. Once those are visible, the probability can be interpreted, challenged, updated, weighted, or rejected without losing the context that made it meaningful.

Key components

Component	Description
Target Event Definition ↗	The numerator-side claim must be named before conditioning begins. A team should know whether the probability concerns disease, default, attack, failure, success, conviction, recovery, or some other target event. Otherwise a precise denominator still produces an ambiguous claim.
Conditioning Context Statement ↗	This is the explicit “given that” clause. It may be a diagnostic signal, a subgroup, a scenario, a legal status, a time period, a model assumption, or a data-selection rule. The clause turns an unconditional probability into a conditioned probability.
Denominator Population ↗	The denominator is the population, cases, probability mass, or model states that remain after the condition is applied. Most practical misuse of conditional probability is denominator drift: using a number estimated in one frame as if it applied to another.
Renormalization Rule ↗	The renormalization rule explains how counts, mass, weights, frequencies, or model probabilities are recalculated inside the conditioned frame. In formal settings this is the familiar P(A | B) structure; in empirical settings it may be a stratified table, model slice, scenario ensemble, or elicited distribution.
Probability Type Label ↗	The same number can be marginal, joint, conditional, posterior, subgroup-conditioned, or scenario-conditioned. The label prevents a conditional probability from being mistaken for an unconditional fact or from being inverted into a different claim.
Frame Compatibility Check ↗	Probabilities should not be compared until their target events, conditioning contexts, denominators, time windows, and sampling rules are compatible. This check is what prevents “apples-to-oranges” probability comparison.
Assumption and Reuse Boundaries ↗	A conditional probability should say what it excludes and where it can travel. A diagnostic estimate from a symptomatic clinic population should not be reused as a public-screening estimate unless the denominator is recalculated.

Common mechanisms

A given-that clause is the simplest mechanism: every probability statement must include its conditioning context. A two-by-two probability table prevents inversion between P(A | B) and P(B | A). A probability tree is useful when conditions arrive sequentially. A reference population note keeps the denominator visible. A stratified rate table supports subgroup-conditioned comparisons. A likelihood-ratio frame distinguishes evidence likelihood from posterior probability. A conditional probability annotation lets model outputs, reports, or data fields carry their context downstream.

Parameters and invariants

Important parameters include the target event, the conditioning event, the denominator size or mass, the time window, the sampling rule, the evidence status, the scenario assumption, the order of nested conditions, and the intended reuse domain.

The invariants are: state the target event, state the condition, preserve the denominator, label the probability type, compare only compatible frames, keep base rates visible when relevant, avoid causal overread, and make the reuse boundary visible.

Target outcomes

A good conditioned probability frame reduces base-rate errors, denominator drift, conditional inversion, and invalid comparisons. It also creates a clean handoff to Bayesian updating, risk weighting, causal analysis, forecast communication, or decision thresholds because each downstream step can see which probability it has actually received.

Tradeoffs

The archetype adds communication overhead. It can also reduce apparent certainty, because once the denominator is visible the number may look more fragile or context-bound. Fine-grained conditioning may improve relevance while shrinking sample size. The goal is not to condition on everything; it is to condition on what the probability claim actually depends on and to say so.

Failure modes

The main failure modes are conditional inversion, denominator drift, base-rate erasure, overconditioned fragility, causal overread, context-stripping in handoff, and incompatible probability comparison. Each failure is a frame error, not merely a calculation error.

Neighbor distinctions

Bayesian Belief Updating revises beliefs over time by combining priors and evidence. This archetype can support Bayesian updating, but it is not the whole update workflow.

Probabilistic Risk Weighting uses probabilities to choose or prioritize action. This archetype makes sure the probabilities being weighted are context-valid.

Coverage Probability Calibration checks whether interval procedures deliver promised coverage. This archetype labels and computes probabilities relative to a given context.

Conditional Independence Boundary Mapping identifies variables that screen a target from its exterior. This archetype states a probability once a condition is taken as given.

Counterfactual Subtraction estimates effects against a constructed absent-intervention baseline. Conditional probability alone does not prove causation.

Examples

In medicine, a positive test result should be reported as probability of disease given the test result and the clinical population, not as test sensitivity alone. In public policy, program outcomes should be compared conditional on eligibility and baseline risk, not simply as raw marginal rates. In security analysis, a threat estimate should say which intelligence holdings and blind spots it is conditioned on. In machine learning, a score should be labeled as conditional on observed features and training-distribution assumptions.

Non-examples

A risk matrix that already receives correctly labeled probabilities is not this archetype; it is downstream risk weighting. A causal estimate from a designed experiment is not primarily this archetype, even though conditional probabilities may appear in the model. A raw intuitive probability with no denominator is also not yet this archetype; the archetype begins when the missing conditioning frame is supplied.

Gap-fill disposition

The pre-draft review found no accepted archetype, alias, variant, component, mechanism, or duplicate-map entry that directly covers conditional_probability as a full conditioning-frame archetype. Related accepted archetypes use conditional probability internally, especially Bayesian belief updating and probabilistic risk weighting, but they do not replace the explicit “given-context plus denominator plus renormalization plus reuse boundary” pattern drafted here.

Common Mechanisms

10 catalogued mechanisms: 9 documented across 4 implementation forms; 1 awaits an authored page and reviewed form classification.

The grouping reflects forms represented among the mechanisms currently documented for this archetype; an absent form is not necessarily an impossible implementation.

Analysis, Modeling & Optimization · 3 mechanisms

• Likelihood-Ratio Frame — Separates the strength of the evidence from the probability of the hypothesis by expressing what a signal says as a ratio that updates a base rate rather than replaces it.
• Stratified Rate Table — Splits an aggregate rate into subgroup rows, each with its own denominator, so subgroup-conditioned probabilities are compared side by side without the marginal hiding them.
• Two-by-Two Probability Table — Lays two binary variables into four joint cells so P(A given B) and P(B given A) are computed from the same grid and can never be confused for each other.

Assessment, Review & Assurance · 1 mechanism

• Frame Compatibility Review — A gate run before two probabilities are compared or pooled, checking that their events, denominators, time windows, and sampling rules are actually commensurable.

Interface, Display & Cue · 1 mechanism

• Given-That Clause — A sentence template that forces every stated probability to name its event and its 'given that' condition in the same breath, so no number is spoken naked.

Representation, Specification & Plan · 4 mechanisms

• Conditional Probability Annotation — Attaches the conditioning frame to a probability value as a machine-readable label so the context travels with the number instead of being stripped downstream.
• Probability Tree — Draws sequential conditions as branching paths, multiplying along each branch so nested 'given that' steps stay in order and the denominator narrows one condition at a time.
• Reference Population Note — A short written note pinning exactly which population a rate is computed over, and where that denominator stops being valid, so the number can't drift onto a different base.
• Scenario Condition Card — One card per named scenario that fixes the full assumption set a forecast is conditioned on, so a scenario-conditioned probability can never be read as unconditional.

Not Yet Form-Classified · 1 mechanism

• Base-Rate Check — Requires the starting prevalence or historical frequency to be considered before vivid evidence is allowed to drive interpretation.

Compression statement

Conditioned probability frame specification turns a naked probability into a scoped claim: identify the event of interest, declare the information context or condition, restrict the denominator to cases compatible with that condition, renormalize probability mass within that restricted frame, and forbid comparisons or updates that silently switch frames.

Canonical formula: P(A | B) = P(A ∩ B) / P(B), with B declared as the operative information context and P(B) > 0

Related Abstractions

Abstractions this archetype builds on — directly (a source ingredient) or as a related pattern. Links follow the typed catalog namespace.

Built directly on (8)

• Bayesian Cue Integration: Multiple noisy cues to one quantity combine by inverse-variance weighting into a lower-variance estimate.
• Bayesian Updating: Update beliefs with evidence.
• Conditional Probability: Re-normalize a probability measure to the information context that is taken as given.
• Context: Surrounding state that selects which content a fixed focal signal carries.
• Encoding Specificity: Retrievability depends on the overlap between features active at encoding and features available at retrieval, because context is co-encoded into the storage key.
• Eventual Realisation of Possibility: In a system subject to many independent trials, every outcome with non-zero per-trial probability eventually occurs, so design posture shifts from per-trial prevention to containment.
• Evidence: A defeasible, provenance-bearing relation between an observable trace and a hypothesis about an unobservable state.
• Probability: Quantifies uncertainty and likelihoods.

Also references 12 related abstractions

• Calibration: Aligning a system's output to a trusted reference by measuring deviation, adjusting to reduce it, and monitoring for drift.
• Causality: Cause-effect relationships.
• Confounding: Hidden variable interference.
• Decision: Committing to one alternative from a set under uncertainty and trade-off, collapsing open deliberation into a chosen path and foreclosing the others.
• Hypothesis Testing (Null vs. Alternative): Null vs alternative evaluation.
• Markov Blanket: The minimal set of variables that, once observed, render a target conditionally independent of everything else — the system's statistical interface with its environment.
• Precision Weighting: Signals about the same latent quantity receive influence in proportion to estimated precision, so more reliable evidence contributes more while context may revise the weights.
• Risk: Exposure to a known distribution of possible outcomes.
• Sampling (Representativeness): Representative subset selection.
• Selection Bias: Skewed sampling.

▸ Show 2 more

Variants

Narrower or domain-specific specializations that share this archetype's core structure. Recognized variants are established; candidate variants are provisional.

Base-Rate Conditioning Frame · domain variant · recognized

A conditioning frame that preserves the relevant reference population before applying evidence or rates.

• Distinct from parent: The parent archetype covers any probability conditioned on given information; this variant emphasizes base-rate and reference-class preservation.
• Use when: A probability claim uses a rate, prevalence, prior frequency, or historical occurrence rate; The risk is that people silently replace the relevant population with a salient subset or anecdotal case; Bayesian updating, diagnosis, screening, or forecasting depends on the denominator being correct.
• Typical domains: medicine healthcare, risk management, forecasting, law governance
• Common mechanisms: base rate table, reference population note, prevalence adjusted risk statement

Subgroup-Conditioned Rate Comparison · scale variant · recognized

A conditioning frame for comparing rates only after subgroup, exposure, or case-mix conditions have been stated.

• Distinct from parent: The parent applies to any conditioned probability; this variant is specialized for comparisons between groups, treatments, sites, or periods.
• Use when: Aggregate and subgroup comparisons may disagree; A rate comparison is vulnerable to case-mix, selection, or confounding differences; Decision-makers need to know whether a claim is marginal, conditional, stratified, or adjusted.
• Typical domains: statistics experimental design, public policy, healthcare quality, organizational metrics
• Common mechanisms: stratified rate table, case mix adjustment note, subgroup slice audit

Evidence-Context Probability Statement · communication variant · recognized

A communication form that states probability together with the evidence, assumptions, and information state it is conditional on.

• Distinct from parent: The parent covers the reasoning architecture; this variant focuses on the statement or reporting artifact.
• Use when: A probability estimate will be communicated to non-specialists or downstream decision-makers; The same number would mean something different under another evidence context; A decision requires knowing whether probability is prior, conditional, posterior, marginal, scenario-specific, or subgroup-specific.
• Typical domains: medicine healthcare, security intelligence, finance, policy analysis
• Common mechanisms: conditional probability annotation, given that clause, probability context card

Near names: Conditional Probability Framing, Conditioning Context Specification, Denominator-Explicit Probability Statement, Information-Conditioned Probability Frame.

Editorial Notes

Problem Classification

Classification: Uncertainty, Evidence & Inference Failure → Probability, Distribution & Risk Calibration

Problem kernel: probability statements omit their conditioning frame

Rationale: Marginal, subgroup, scenario, and evidence-conditioned probabilities are compared as though they referred to the same information state.

Independent corroboration: The earliest necessary condition in the frozen evidence is: Probability claims are made, compared, or used in decisions without declaring what information, subgroup, scenario, evidence state, or assumption set is being taken as given. That is a probability distribution and risk calibration problem because Probability, uncertainty intervals, tails, multiplicity, and variability are interpreted under hidden frames or assumptions that misstate risk.

Review outcome: Independent reviewer agreement; high confidence.