Antecedent 2.4.0

Changes since Antecedent 2.3.0. The release connects measurements, fitted mechanisms, external evidence, transported claims, adaptive experiments and decisions through checked composition: each handoff states what it carries, what it cannot supply, and what it refuses. Every new interval route either carries coverage records at its own coordinate or is labelled derived and unmeasured. This page records each new cell with its inference status, the adjacent requests that are refused, and the artifact compatibility classes.

Consumer workflows

If you need to…	Start here	What 2.4 adds	Inference status
Correct a measured exposure or outcome, or use an external latent posterior	Measurement transport, interface contracts	Classical error, repeated measurements, binary misclassification, categorical confusion matrices, provider latent posteriors, one measurement-model contract	point-only
Compose fitted mechanisms and propagate dependent uncertainty	Composition, composed uncertainty	Acyclic joint law under intervention; linear moments, aligned nonlinear functionals, known-weight mixtures; spread and exceedance of the composition	statistical interval for linear moments; derived-unmeasured otherwise
Pool, borrow from or revise evidence without double counting	External evidence, assumption range and revalidation	Row-level overlap, minimum-variance and random-effects pooling, hierarchical borrowing, a correlation range, revalidation after a source changes	statistical interval for fixed-effect and random-effects pooling; point-only otherwise; assumption range for unknown correlation
Transport counterfactuals, mechanisms and cohort effects	Transport and counterfactuals, mechanism transport, discovered structure	Nested and conditional counterfactual transport, sampled transport, transported mechanisms, class-of-structures transport, provider-sourced evidence, cohort-effect decisions	statistical interval for sampled counterfactual transport, cross-world quantities and the native cohort mean; otherwise point-only, exact-finite or structural envelope
Decide or invert on composed, cohort or transported sources	Inverse and decision reach, decisions	Sources state before execution what they supply; their identity and transport receipt travel into the result	follows the source
Judge how reliable a sampled decision or ranking is	Decision reliability	Monte Carlo error and effective sample size, ranking probabilities, selection-aware evaluation	statistical interval for ranking probability, utility, confidence set and selection-corrected value; derived-unmeasured otherwise
Run and analyse an adaptive experiment	Adaptive procedures	Analysis contract, procedure matrix, posterior, decision-theoretic stopping, adaptive weighting, repeated measures, anytime-valid and group-sequential readings	statistical interval for the measured Thompson, epsilon-greedy repeated-measures, anytime-valid and group-sequential coordinates; other guarantees closed
Assess a predictive law, or condition on an observation	Capabilities	Proper scores, score differences, calibration diagnostics, drift; observation-conditioned probabilities and evidence requirements	statistical interval for the score and the count-fitted observation probability; point-only otherwise
Re-run discovery with a changed setting	Selective recalculation	Stage reuse under a changed threshold, multiplicity rule, background knowledge, orientation budget, lag window or CI test	none (reuse accounting only)
Estimate an effect after discovering the graph	Selective discovery, discovered structure	Discovery on one half of a seeded split, a resolved structure accepted, the effect adjusted for the treatment's parents on the other half; the interval is attested only from a coverage record of the whole method	statistical interval for the split-discovery effect; per-dataset recovery stays closed
Call a provider you do not own	Provider-sourced transport	Law, posterior and nested callbacks with replay policy, portable results and re-derived identities	none (claims, not native facts)

Bayesian counterparts of the vector, categorical, dose, mediation, latent-class, penalized and clustered score, finite-law decision, recovery and transport routes return a posterior mean as a point. Their draws are never offered as an interval.

Inference status

Cells outside the table below carry one of point_only (exact-finite or plug-in point), assumption_range (value over a declared premise range) or none (identification, inspection or a closed adjacent claim). Derived uncertainty is exact under its stated model and carries derived-unmeasured standing. One hundred and three coverage records are new in this release, each measured at 400 and rechecked at 2,000 replicates over a three-point sample-size grid, with borderline cells re-measured at 8,000 and 20,000.

Route	Measured coordinate	Sample-size range	Coverage (nominal level)
Sampled counterfactual transport interval	three finite designs	250 to 4,000 units	0.94 to 0.96
Whole-row bias-corrected recovery interval	binary missingness	1,000 to 4,000 rows	0.93 to 0.96
Checked temporal studentized and basic effect intervals	whole-unit resampling	64 to 256 units	0.93 to 0.96
Pooled evidence interval	fixed effect and Hartung-Knapp random effects, 5 and 10 studies	75 to 200 per study	0.93 to 0.96
Cohort mean effect from a native linear fit	HC3 with a t quantile, 3 to 5 covariates	200 to 800 rows	0.90 to 0.97
Count-fitted observation probability	Dirichlet posterior, confounded and measured-signal designs	150 to 600 rows	0.95 to 0.96
Ranking probability, utility, confidence-set and selection-corrected intervals	2, 8 and 64 candidates, shared and separate streams	500 to 2,000 draws	0.94 to 0.96
Predictive score interval	iid, cluster and serial dependence	400, 800 and 1,600 cases upward	0.93 to 0.97
Bridged claim interval	trial-level prediction and measurement correction	15 to 40 trials; 100 to 400 units	0.93 to 0.96
Adaptive experiments	Thompson posterior (prior-predictive and large gap), epsilon-greedy repeated measures from 300 units, anytime-valid sequence	100 to 1,200 units	0.93 to 0.97
Group-sequential repeated interval	three looks, null to large effect	120 to 480 units	0.94 to 0.96
Composed linear-moment interval	shared rows, independent samples, aligned bootstrap	200 to 800 rows	0.94 to 0.96
Cross-world intervals	independent sample from 60 units, fixed units	60 to 240 units	0.94 to 0.96
Split-discovery downstream effect	discover on one half, adjust on the other	300 to 1,200 analysis rows (600 to 2,400 in the table)	0.95 to 0.96

The limits are recorded as named boundaries, not claimed: the Thompson posterior at exactly equal arms (0.92 to 0.94), a weighted Wald interval read at a posterior-probability stopping look (0.73 to 0.95 depending on exploration), the checked percentile effect interval (0.93 to 0.945), the serial CRPS mean interval (0.945), the cohort interval at 200 rows with four covariates (conservative, 0.918) and split discovery under weak edges, cancelling edges or reuse of the discovery rows. Existing measured scalar intervals from 2.3 keep their records and scope; none of them extends to a result outside its coordinate by inheritance, and a calibration record is never carried to a new target or composition.

Refused adjacent requests

Request	Reason	Remedy
A recovery, false-discovery or error-control statement about the graph discovered on one dataset, or an interval for an estimate computed on the rows discovery used	Faithfulness and causal sufficiency are not testable from the data discovery ran on; reused rows have no measured construction	Report the structure with its unresolved edges; split the rows and use the sample-split effect
An interval for hierarchical, accounting-only or dependence-envelope pooled results, a nonlinear composed quantity, a bridged claim with no source standard error, a cohort mean from a caller-built export, an adaptive result read at a stopping look, a declared-model observation probability or a revalidated claim	No record at that coordinate, or no sampling distribution exists	Use the point or the assumption range; supply the missing input; measure the coordinate
Ranking probability, optimality set or selected-value interval from a point-only producer or independently drawn candidates	Candidates must come from one joint law	Keep per-draw utilities from one joint law
Source-spread, Monte Carlo and selection uncertainty as one combined number	Distinct meanings are not added	Report them separately
A probability, quantile, nonlinear utility or joint law from a cohort mean or a point source	A mean supplies no law	Supply draws or a finite table
A mechanism behind a selection arrow, or a transport with no declared uncertainty	The premise is untested	Refit in the target; declare the variance and covariance
A cross-world value in another population, with sampled abduction, or with an interval	Mechanisms are declared for one population	Declare the population; use the point
A cyclic composition, or inherited calibration from an input without a matching upstream record	Calibration is never inherited	Break the cycle; re-measure
Fixed-sample inference on an adaptive design, posterior under value-dependent loss, coverage of an adaptive result	Paradigms are not interchangeable	Model the loss; use the adaptive-weighting or anytime-valid paradigm
Unknown overlap treated as disjoint, or independence declared where overlap is unknown	Overlap is decided from row-level identity	Register the sources in the evidence ledger
A score interval, significance level, callback-only law, or causal validity from a predictive score	Scores describe realized observations only	Use the score as a diagnostic
A callback supplying a posterior guarantee, native trust, identification licence or joint law	Callback results are claims	Verify the object; use a native route

What these routes do not do lists each package's most likely refused request and remedy. A refusal carries a registered reason code, a typed detail and a remedy.

Artifact compatibility

Change	Class	Migration
Decision contracts and results, external claims, composition bundles, inverse-query and robust decision results: new identity domains	Requires regeneration	Identity domains
Design ranking request and signal fingerprints	Inspectable only	Fingerprints
Design ranking verification labels, overlap decisions and Monte Carlo streams	Migrated	Verification and overlap
Recovery, mixed-source and study-plan derivations and rule-version spellings	Requires regeneration (wire inspectable)	Recovery identity
Composite recalculation decision stage	Migrated	Stage identity
Joint-law and decision-source artifacts	Byte-compatible	Identity domains
Python API: retained studies, read models, bytes-first export	Source-incompatible renames	Python migration

Scope of the measured claims

Every uncertainty route that carries a calibrated interval in this release has its coverage records at the coordinate it is licensed for, and refuses with a typed reason outside that coordinate. No result is issued as calibrated without a coverage record at its own coordinate.

Python coverage

Every public facade entry has a Python entry except antecedent::analysis::api::posterior_predictive_predictor. The predictor takes the Rust-typed prepared Bayesian problem and posterior, which Python holds no handle to; replicate_laws covers replicates Python already has.

How to read the claims

Term	Meaning
`point_only`	An exact or plug-in point estimate. No interval of any kind is claimed.
`nominal`	An interval whose coverage was never measured at its own coordinate (`estimator_grid_not_measured`). Unmeasured diagnostics grant no coverage license.
`calibrated`	An interval with a coverage record measured at its own coordinate; shown to readers as a statistical interval. A cell is described this way only once its records exist.
`structural_envelope`	The range of an answer over explicitly supplied structural scenarios. Not a confidence interval.
`assumption_range`	The range of an answer over a declared sensitivity set. Not a confidence interval.
`none`	No numerical claim: identification or inspection only.
statistical interval	The reader-facing name of a `calibrated` interval; it appears only beside a measured coverage record.

Claims by cell

vector treatment.bayesian

Claim: point_only

Question: Can the posterior mean of the joint treatment coefficients be consumed without treating draws as an interval?
Scope: No graph is identified. The caller owns the shared adjustment set.
Guarantee: joint_coefficient_posterior_mean_point_only.
Bounds: none declared.

categorical treatment.bayesian

Claim: point_only

Question: Can the posterior mean of the dummy-coded categorical coefficients be consumed without an interval?
Scope: No graph is identified.
Guarantee: categorical_coefficient_posterior_mean_point_only.
Bounds: none declared.

quadratic dose.bayesian

Claim: point_only

Question: Can the posterior mean of a local quadratic dose response be consumed as a point?
Scope: No graph is identified.
Guarantee: local_quadratic_posterior_mean_point_only.
Bounds: none declared.

dose grid functional.bayesian

Claim: point_only

Question: Can the posterior mean of a dose-grid local quadratic be consumed as a point?
Scope: No graph is identified.
Guarantee: dose_grid_posterior_mean_point_only.
Bounds: none declared.

functionals.bayesian

Claim: point_only

Question: Can the functional of a Dirichlet posterior mean of the finite law be consumed as a point?
Scope: No graph is identified. The caller owns the decision contract.
Guarantee: decision_functional_of_dirichlet_mean_point_only.
Bounds: none declared.

functional inverse.bayesian

Claim: point_only

Question: Can an inverse of the functional of a Dirichlet posterior mean be consumed as a point?
Scope: No graph is identified.
Guarantee: functional_inverse_of_dirichlet_mean_point_only.
Bounds: none declared.

finite action inverse.bayesian

Claim: point_only

Question: Can a finite-action inverse of a Dirichlet posterior mean be consumed as a point?
Scope: No graph is identified.
Guarantee: finite_action_inverse_of_dirichlet_mean_point_only.
Bounds: none declared.

binary observation.bayesian

Claim: point_only

Question: Can binary observation recovery of a Dirichlet posterior mean be consumed as a point?
Scope: The observation graph the frequentist recovery already uses.
Guarantee: binary_recovery_of_dirichlet_mean_point_only.
Bounds: none declared.

chain observation.bayesian

Claim: point_only

Question: Can chain observation recovery of a Dirichlet posterior mean be consumed as a point?
Scope: The observation graph the frequentist chain recovery already uses.
Guarantee: chain_recovery_of_dirichlet_mean_point_only.
Bounds: none declared.

penalized propensity aipw.bayesian

Claim: point_only

Question: Can the posterior mean of a penalized AIPW score contrast be consumed as a point?
Scope: The back-door adjustment DAG of the frequentist penalized AIPW route.
Guarantee: penalized_score_mean_posterior_point_only.
Bounds: none declared.

clustered dml aipw.bayesian

Claim: point_only

Question: Can the posterior mean of a clustered DML score contrast be consumed as a point?
Scope: The back-door adjustment DAG of the frequentist clustered DML route.
Guarantee: clustered_score_mean_posterior_point_only.
Bounds: none declared.

batch retarget covariance.bayesian

Claim: point_only

Question: Can the posterior mean of the cross-fitted score a batch retarget reads be consumed as a point?
Scope: The back-door adjustment DAG of the score table the batch reads.
Guarantee: batch_score_mean_posterior_point_only.
Bounds: none declared.

multi source limited experiment.bayesian

Claim: point_only

Question: Can a Dirichlet posterior mean of the supplied finite laws be pushed through the multi-source identifier?
Scope: The selection diagram of transport.multi_source_limited_experiment.
Guarantee: multi_source_dirichlet_mean_point_only.
Bounds: none declared.

two step temporal.bayesian

Claim: point_only

Question: Can a Dirichlet posterior mean of the supplied finite laws be pushed through the two-step temporal identifier?
Scope: The existing fixed two-slice selection diagram. No TemporalAdmg.
Guarantee: two_step_dirichlet_mean_point_only.
Bounds: none declared.

admg conditional.bayesian

Claim: point_only

Question: Can a Dirichlet posterior mean of the supplied finite laws be pushed through the conditional ADMG identifier?
Scope: The selection diagram of transport.admg_conditional. No TemporalAdmg.
Guarantee: admg_conditional_dirichlet_mean_point_only.
Bounds: none declared.

shared scenario covariance.bayesian

Claim: point_only

Question: Can the posterior mean of the shared scenario covariance be consumed as a point?
Scope: The scenario set the frequentist covariance already reads. The envelope stays structural.
Guarantee: scenario_covariance_posterior_mean_point_only.
Bounds: none declared.

uncertain initial state.bayesian

Claim: point_only

Question: Can the posterior mean of the uncertain initial-state mixture be consumed as a point?
Scope: The existing fixed two-slice selection diagram.
Guarantee: initial_state_mixture_posterior_mean_point_only.
Bounds: none declared.

new period refresh.bayesian

Claim: point_only

Question: Can the posterior mean of a refreshed period response be consumed as a point?
Scope: The existing fixed two-slice graph when the window advances.
Guarantee: refreshed_period_posterior_mean_point_only.
Bounds: none declared.

effect constancy.bayesian

Claim: point_only

Question: Can the posterior mean of an effect-constancy contrast be consumed as a point?
Scope: The declared partition. This cell identifies no graph.
Guarantee: constancy_contrast_posterior_mean_point_only.
Bounds: none declared.

nonlinear mediation.bayesian

Claim: point_only

Question: Can the posterior mean of the mediator and outcome coefficients be consumed as a point?
Scope: The same declared mediation premises as the frequentist record. No new graph.
Guarantee: nonlinear_mediation_coefficient_posterior_mean_point_only.
Bounds: none declared.

latent class effects.bayesian

Claim: point_only

Question: Can the posterior mean of each latent class's treatment effect be consumed as a point?
Scope: The same two-class mixture as the frequentist record. No new graph.
Guarantee: latent_class_effect_posterior_mean_point_only.
Bounds: none declared.

mechanism discrepancy.bayesian

Claim: point_only

Question: Can the posterior mean of a population's mechanism coefficients be consumed as a point?
Scope: The same comparable linear mechanism as the frequentist record. No new graph.
Guarantee: mechanism_coefficient_posterior_mean_point_only.
Bounds: none declared.

marginal structural model.bayesian

Claim: point_only

Question: Can a Dirichlet posterior mean of the finite stratum laws, pushed through the marginal structural model at Lambda 1, be consumed as a point?
Scope: No graph is identified. The caller supplies the same finite strata as the frequentist record.
Guarantee: msm_dirichlet_identified_ate_point_only.
Bounds: none declared.

learned continuous trial.bayesian

Claim: point_only

Question: Can the posterior mean of the learned continuous trial's score table be consumed as a point?
Scope: The same trial-to-target certificate as the frequentist record. No new graph.
Guarantee: learned_continuous_score_posterior_mean_point_only.
Bounds: none declared.

smoothed dose response.bayesian

Claim: point_only

Question: Can the posterior mean of each smoothed-dose score table be consumed as a point?
Scope: The same dose-response certificate as the frequentist record. No new graph.
Guarantee: smoothed_dose_score_posterior_mean_point_only.
Bounds: none declared.

acyclic joint model

Claim: point_only

Question: Do independently fitted conditional mechanisms establish one coherent acyclic joint law, and what does that law say under intervention, with parameter uncertainty propagated only under a declared dependence?
Scope: Static acyclic directed graph over named variables; linear-Gaussian additive-noise mechanisms with declared unit-variance Gaussian shared latents and loadings, or finite-discrete conditional tables; one family per composition.
Guarantee: exact_given_declared_mechanisms_point_only.
Bounds: max_mechanisms = 64, max_parents_per_mechanism = 16, max_latents = 32, max_finite_states = 1048576, max_sample_cells = 67108864.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

overlap accounting

Claim: none

Question: Which observations do two pieces of evidence share, and how many does a combined result count once?
Scope: No graph. Evidence sources are observation sets, experiments, priors, fitted models and external claims with declared row identities and lineage.
Guarantee: exact_set_arithmetic_over_declared_observation_identities_no_inference.
Bounds: none declared.

statistical pooling

Claim: point_only

Question: What is the minimum-variance linear pooled estimate of study summaries with a stated covariance, and may these sources be pooled at all?
Scope: No graph. Summary data: one estimate per source and a covariance.
Guarantee: gls_point_estimate_under_declared_covariance_point_only_no_interval.
Bounds: none declared.

random effects

Claim: point_only

Question: What is the between-study variance and the random-effects pooled mean of independent study summaries?
Scope: No graph. Summary data: one estimate and variance per source.
Guarantee: random_effects_point_estimate_conditional_variance_no_interval.
Bounds: none declared.

hierarchical borrowing

Claim: point_only

Question: What do the studies' partially pooled effects and the common mean look like under a normal-normal hierarchical model with a declared prior on the between-study standard deviation?
Scope: No graph. Summary data with a declared exchangeable-study model.
Guarantee: hierarchical_normal_posterior_mean_variance_exact_for_declared_finite_prior_point_only.
Bounds: none declared.

transport provenance

Claim: none

Question: Does a claim carry its transport identity so that it is never counted as native target data?
Scope: The selection diagram is carried as a digest; no graph is interpreted by the ledger.
Guarantee: transport_identity_bookkeeping_no_inference.
Bounds: none declared.

evidence reuse

Claim: none

Question: May a result be cited again in a second analysis without its observations being counted twice?
Scope: No graph.
Guarantee: reuse_counts_cited_result_once_no_inference.
Bounds: none declared.

repeated observation use

Claim: none

Question: May the same observations serve a later stage of one workflow, and is the later result independent confirmation?
Scope: No graph.
Guarantee: repeated_use_declared_flagged_counted_once_no_inference.
Bounds: none declared.

revision

Claim: none

Question: What happens to dependent claims when a source is updated, replaced, withdrawn or superseded?
Scope: No causal graph; a lineage and result dependence structure only.
Guarantee: revalidation_classification_exact_over_declared_lineage_no_inference.
Bounds: none declared.

dependence range

Claim: assumption_range

Question: How far can the pooled estimate and its variance move when the studies' sampling correlation is unknown but lies in a declared interval?
Scope: No graph. Summary data with one declared common correlation interval.
Guarantee: exact_envelope_over_declared_correlation_range_assumption_range_only.
Bounds: none declared.
Not an interval: the assumption range is never a confidence interval.

pooled interval

Claim: calibrated

Question: Can an interval, posterior or coverage statement be reported for pooled or borrowed evidence?
Scope: No graph.
Guarantee: pooled_mean_interval_measured_fixed_effect_wald_and_random_effects_hksj_point_otherwise.
Bounds: none declared.

model contract

Claim: point_only

Question: Can a measured column be related to the latent scientific quantity it is offered for, bound to a population, regime, arm and representation version, with differential error declared and proxy-as-construct refused?
Scope: Not applicable: the contract binds a measured column to a construct and imposes no graph class; identification of any downstream estimand belongs to the identification layer.
Guarantee: typed_measurement_semantics_only_no_identification_or_coverage.
Bounds: max_artifact_bytes = 8388608.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

classical continuous error

Claim: point_only

Question: With additive classical measurement error of known or validated variance in a continuous regressor, what is the corrected linear adjusted effect and its standard error including the error-variance uncertainty?
Scope: Not applicable to the correction; the corrected coefficient is a causal effect only where the identification layer licenses the adjustment set and the linear model.
Guarantee: corrected_linear_coefficient_point_with_uncalibrated_delta_method_se.
Bounds: max_rows = 67108864, max_covariates = 64.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

repeated measurements

Claim: point_only

Question: With replicated error-prone measurements of a continuous regressor, what are the within/between variance components and ICC, and what is the corrected linear adjusted effect with the replicate-variance uncertainty propagated?
Scope: Not applicable to the correction; causal reading requires identification of the adjustment set.
Guarantee: replicate_variance_components_and_corrected_linear_coefficient_point_with_uncalibrated_se.
Bounds: max_rows = 67108864, max_covariates = 64.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

binary misclassification

Claim: point_only

Question: With a binary outcome misclassified at known or validated sensitivity and specificity, what are the corrected prevalence and risk difference, with the sensitivity/specificity uncertainty propagated and arm-specific error kept separate?
Scope: Not applicable; the risk difference is a causal contrast only where exposure assignment is identified.
Guarantee: corrected_prevalence_and_risk_difference_point_with_uncalibrated_delta_method_se.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

probabilistic categorical

Claim: point_only

Question: With categorical labels assigned under a declared confusion matrix, what are the corrected category proportions, expected counts and category outcome means, with confusion-matrix uncertainty propagated?
Scope: Not applicable; category outcome contrasts are associations unless identification licenses a causal reading.
Guarantee: corrected_category_summaries_point_with_uncalibrated_delta_method_se.
Bounds: max_categories = 16.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

external latent posterior

Claim: point_only

Question: Given a latent-variable posterior supplied by an external provider, what is the linear outcome effect of the latent regressor, how far is the provider trusted, and is the provider's own uncertainty propagated?
Scope: Not applicable; identification of the outcome model is separate.
Guarantee: externally_supplied_posterior_linear_coefficient_point_with_conditional_se.
Bounds: max_rows = 67108864, max_covariates = 64, max_imputations = 10000.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

transport general

Claim: point_only

Question: What is a nested or conditional counterfactual probability in a target population, given labelled observational and interventional laws from the target and from source populations whose mechanisms may differ at declared variables?
Scope: Static finite-discrete acyclic semi-Markovian ADMG over at most 8 observed variables with at most 4 declared levels each; one selection diagram per source over the shared graph.
Guarantee: sound_incomplete.
Bounds: max_observed = 8, max_levels = 4, max_worlds = 6, max_atoms = 16, max_network_nodes = 24, max_sources = 4, max_controllable_per_source = 4, max_regimes = 64, max_sum_cells = 65536, max_operations = 400000, max_depth = 256.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

transport sampled

Claim: point_only

Question: What is the estimated value of a nested or conditional counterfactual probability in a target population when the observational and interventional laws of the target and of the sources are known only through sampled counts, and what does the sample license about it?
Scope: Static finite-discrete acyclic semi-Markovian ADMG over at most 8 observed variables with at most 4 declared levels each; the selection diagram and bounds of counterfactual.transport_general. Continuous, cyclic, temporal and selected-sample designs are separate rows.
Guarantee: plug_in_point_estimate_standard_error_uncalibrated_interval_closed.
Bounds: max_tables = 64, max_table_cells = 65536, max_envelope_evaluations = 65536.

transport sampled interval

Claim: calibrated

Question: Can an interval, posterior or coverage statement be reported for a sampled counterfactual transport estimate?
Scope: As counterfactual.transport_sampled.
Guarantee: delta_wald_interval_measured_named_queries_independent_units_refused_unseen_cells.
Bounds: none declared.

mean contrast

Claim: point_only

Question: What is the weighted mean treatment contrast of a retained fitted effect model over a declared cohort, as one native quantity a decision can use under an affine utility?
Scope: Not applicable beyond the parent analysis: the parent claim's identification (adjustment by the model's features) is retained, not re-derived.
Guarantee: weighted_cohort_mean_of_retained_contrast_predictions_affine_expectation_point_only.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

declared shift

Claim: point_only

Question: Can a fitted effect model decide for a cohort drawn from a different population when the conditional contrast is declared invariant across the two populations and every cohort row lies in the fitted support?
Scope: A selection diagram declared by the caller: the invariant effect modifiers are exactly the model's features and the selection node enters only them.
Guarantee: declared_invariance_cohort_mean_point_only_extrapolative_label.
Bounds: none declared.

interval

Claim: calibrated

Question: Can an interval be reported for the declared-cohort mean of a linear conditional treatment effect, and under which fit does it carry measured coverage?
Scope: As cohort_decision.mean_contrast.
Guarantee: student_t_hc3_cohort_mean_interval_measured_native_linear_effect_fit_declared_export_point_only.
Bounds: none declared.

law valued request

Claim: none

Question: Can a cohort-mean contrast answer a probability, a quantile, a nonlinear utility or a joint law?
Scope: Not applicable.
Guarantee: law_valued_requests_refused_a_cohort_mean_determines_no_law_closed_by_logic.
Bounds: none declared.

sampled decision

Claim: point_only

Question: How reliable is the numerical value of a decision computed from draws of a supplied law: what are the Monte Carlo standard error and effective sample size of a utility, a utility difference, a threshold probability or a risk measure, and what is the spread of the utility under the source law, reported separately?
Scope: No graph. A finite candidate set of per-draw utilities supplied by a producer that declares the draws source, coupling and serial dependence.
Guarantee: numerical_standard_error_and_effective_sample_size_uncalibrated_no_interval.
Bounds: max_candidates = 65536, max_cells = 67108864.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

ranking statement

Claim: point_only

Question: Over a finite candidate set, with what Monte Carlo probability under the supplied law is an action best, in the top k, or better than a baseline?
Scope: No graph. A finite candidate set of per-draw utilities supplied by a producer that declares the draws source, coupling and serial dependence.
Guarantee: monte_carlo_ranking_diagnostic_uncalibrated_no_interval.
Bounds: max_ranking_cells = 16777216.

selection aware evaluation

Claim: point_only

Question: When the reported decision was chosen among many noisy candidates, how far does its in-sample value overstate its value, and what does an independent re-evaluation of the selected decision give?
Scope: No graph. A finite candidate set of per-draw utilities supplied by a producer that declares the draws source, coupling and serial dependence.
Guarantee: winners_curse_bound_and_held_out_reevaluation_point_only_no_interval.
Bounds: none declared.

ranking probability interval

Claim: calibrated

Question: Can a ranking or optimality probability, a confidence or credible set for the best action, or an interval for a selected decision's value be published as a calibrated claim?
Scope: No graph. A finite candidate set of per-draw utilities supplied by a producer that declares the draws source, coupling and serial dependence.
Guarantee: published_probability_and_interval_routes_resolved_from_current_coverage_record_at_the_coordinate_refused_elsewhere.
Bounds: none declared.

combined uncertainty

Claim: none

Question: Can source-distribution spread, Monte Carlo error and selection-induced uncertainty be reported as one combined number?
Scope: No graph. A finite candidate set of per-draw utilities supplied by a producer that declares the draws source, coupling and serial dependence.
Guarantee: combined_uncertainty_route_closed.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: uncertainty_components_not_combined).

proper scores

Claim: point_only

Question: What do the log score, the CRPS, the Brier score and the energy score of a predictive law say about realized observations, with exact values for known laws and a recomputing consumer?
Scope: No graph. Predictive laws and realized observations; the assessment is model-independent.
Guarantee: exact_closed_form_or_finite_law_scores_point_only_no_interval.
Bounds: max_cases = 1000000, max_predictors = 32, max_law_atoms = 65536, max_joint_dimension = 64.

score differences

Claim: point_only

Question: Does one named predictor score better than another on the same realized observations, with the dependence from the shared observations and the cases stated and not assumed independent?
Scope: No graph. Predictive laws and realized observations; the assessment is model-independent.
Guarantee: paired_difference_with_declared_dependence_standard_error_diagnostic_no_interval.
Bounds: none declared.

calibration diagnostics

Claim: point_only

Question: How did a predictive law behave on the realized observations: is its PIT uniform, are its forecasts reliable, do its prediction intervals cover at their level, and how sharp is it?
Scope: No graph. Predictive laws and realized observations; the assessment is model-independent.
Guarantee: labelled_unmeasured_diagnostics_no_calibration_guarantee_no_interval.
Bounds: none declared.

overlap check

Claim: none

Question: Were the realized observations also used to fit or tune the assessed law, decided from row-level identities and not from a label?
Scope: No graph. Predictive laws and realized observations; the assessment is model-independent.
Guarantee: exact_set_arithmetic_over_declared_observation_identities_no_inference.
Bounds: none declared.

distribution drift

Claim: point_only

Question: Did the distribution of a law's probabilistic errors move between two windows of the assessed observations?
Scope: No graph. Predictive laws and realized observations; the assessment is model-independent.
Guarantee: labelled_unmeasured_drift_diagnostic_no_significance_level.
Bounds: none declared.

score interval

Claim: calibrated

Question: Can an interval or significance level be reported for a score or a score difference?
Scope: No graph. Predictive laws and realized observations; the assessment is model-independent.
Guarantee: expected_score_interval_resolved_from_current_coverage_record_unscoped_and_unmeasured_requests_refused.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

law callback

Claim: none

Question: Can a predictive law that exists only as a callback be scored?
Scope: No graph. Predictive laws and realized observations; the assessment is model-independent.
Guarantee: callback_law_route_closed_not_built.
Bounds: none declared.

causal validity

Claim: none

Question: Does a good predictive score establish causal validity, or the probability that a scientific claim holds?
Scope: No graph. Predictive laws and realized observations; the assessment is model-independent.
Guarantee: score_not_causal_validity_route_closed.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: route_not_supported).

finite law conditioning

Claim: point_only

Question: Given a licensed causal model over finite variables, what is the probability of an outcome given an observed signal, how does it differ from the probability under an intervention that sets the signal, and what changes when the signal is a measured value of a latent construct?
Scope: Static acyclic directed graph over finite-discrete variables compiled from declared conditional tables; one executable joint law per composition, enumerated within the finite-state bound.
Guarantee: exact_given_declared_model_or_labelled_point_value_of_a_fitted_model_no_interval.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

posterior update

Claim: point_only

Question: After new evidence, what is the posterior over a declared finite hypothesis space or a declared Beta-Bernoulli model, and what is the probability of the next observation?
Scope: Not applicable: a declared finite hypothesis space with a likelihood table, or a one-parameter conjugate model.
Guarantee: exact_given_the_declared_prior_and_likelihood_no_interval.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

evidence requirement

Claim: point_only

Question: How much evidence is needed to move a posterior or decision probability past a declared level, and is that level reachable at all?
Scope: Not applicable: a declared binary hypothesis with a declared signal, or a Beta-Bernoulli predictive probability.
Guarantee: exact_minimum_given_the_declared_signal_or_a_refusal_stating_the_reachable_extremum.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

calibrated probability

Claim: calibrated

Question: Can an interval or a calibration claim be reported for a posterior or decision probability computed from an observation-conditioned query, and for which construction is it earned?
Scope: As observation_query.finite_law_conditioning.
Guarantee: declared_model_has_no_sampling_interval_count_fitted_posterior_interval_measured.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

typed bridges

Claim: point_only

Question: Under which declared assumptions does a claim about one scientific quantity answer a question about another (a latent construct and its measured outcome, a stated preference and realized behavior, a surrogate and a terminal outcome, a predictive quantity and a causal quantity, a model parameter and an intervention response, a scientific endpoint and a decision utility), and which substitutions refuse?
Scope: Not applicable to the bridge itself. The predictive-to-causal bridge consumes an identification (adjustment set, exchangeability, positivity, consistency) established or declared by the identification layer; the bridge records it as a named assumption and does not check it.
Guarantee: declared_assumption_bridges_point_only_no_interval_or_law.
Bounds: max_assumptions = 64, max_validation_trials = 4096.

law substitution

Claim: none

Question: Can a bridged mean stand in where a predictive law, a posterior, an identified set or a joint outcome law is required, or the reverse?
Scope: Not applicable.
Guarantee: value_kind_substitution_refused_no_interconversion.
Bounds: none declared.

interval

Claim: calibrated

Question: Can an interval, posterior or coverage statement be reported for a bridged claim?
Scope: As scientific_bridge.typed_bridges.
Guarantee: bridged_value_interval_from_declared_source_standard_error_resolved_from_current_coverage_record_unscoped_posterior_and_joint_law_refused.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

policy evaluation

Claim: point_only

Question: What is the value of a caller-supplied contingent policy over a finite horizon of action, observation, belief update and action, with delayed outcomes, a resource budget, stopping boundaries and a terminal utility, and how does acting now compare with waiting, collecting evidence or choosing differently after an observation?
Scope: A finite-horizon dynamic finite structural model with at most 32 latent states, 16 actions and 12 decision times; the information set of a decision is the observed history of action and observation pairs.
Guarantee: fixed_policy_value_exact_enumeration_and_importance_weighted_estimate_point_only.
Bounds: max_states = 32, max_actions = 16, max_horizon = 12, max_policy_entries = 65536, max_logged_trajectories = 150000, max_rollouts = 10000000.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

analysis contract

Claim: none

Question: Is the retained analysis contract of an adaptive experiment complete and consistent enough to license any inference, and which element is missing when it is not?
Scope: Not applicable.
Guarantee: semantic_contract_exact.
Bounds: none declared.

procedure matrix

Claim: none

Question: For a given assignment policy, outcome structure, monitoring strategy and stopping rule, does the engine license inference, refuse it, or carry it forward with an owner?
Scope: Not applicable.
Guarantee: semantic_contract_exact.
Bounds: none declared.

bayesian posterior

Claim: point_only

Question: After a Thompson-sampling trial with interim looks, what is each arm's posterior and which arm is best, valid despite the adaptive assignment?
Scope: Not applicable; the arm means are causal contrasts only where the randomization is the identifying assumption.
Guarantee: exact_posterior_given_declared_prior_point_only.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

decision theoretic stopping

Claim: point_only

Question: At an interim look of a Thompson-sampling trial, does the declared posterior expected-loss rule say stop, and which arm would be chosen?
Scope: Not applicable.
Guarantee: exact_decision_given_declared_loss_and_prior_point_only.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

adaptive weighting

Claim: point_only

Question: From an epsilon-greedy trial with interim looks, what is an arm's mean or the contrast of two arms, estimated so that the adaptive assignment does not bias it?
Scope: Not applicable; the contrast is causal where assignment given history is the identifying assumption, which the retained probabilities encode.
Guarantee: adaptively_weighted_point_with_uncalibrated_martingale_variance.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

repeated measures weighting

Claim: point_only

Question: From an adaptive trial of subjects each measured repeatedly, what is the contrast of arm means with the subject, not the measurement, as the unit?
Scope: Not applicable.
Guarantee: subject_level_adaptively_weighted_point_with_uncalibrated_martingale_variance.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

anytime valid sequence

Claim: point_only

Question: For bounded outcomes under adaptive assignment and data-dependent stopping, what time-uniform confidence sequence does each arm's mean have?
Scope: Not applicable.
Guarantee: time_uniform_sequence_theorem_unmeasured_as_a_procedure.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

group sequential

Claim: calibrated

Question: For a randomized trial with fixed allocation and pre-specified interim looks, what efficacy boundaries spend the declared alpha, and did the observed statistic cross one?
Scope: Not applicable.
Guarantee: alpha_spending_boundaries_point_and_repeated_confidence_interval_measured_stagewise_adjustment_closed.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

guarantees closed

Claim: calibrated

Question: Can an interval, credible interval, error rate or coverage statement be reported for an adaptive procedure's result?
Scope: As the adaptive_inference records.
Guarantee: measured_interval_for_three_procedure_cells_fixed_sample_and_stopped_wald_closed.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

claim revalidation

Claim: none

Question: When an upstream object of a composed analysis changes (data, evidence, a model or representation version, a structural derivation, a target population, a selection diagram, support, provider trust or calibration), which downstream claims are still valid, which must be recomputed and which are no longer licensed, and which upstream claims, bridges, shared data, versions, unsupported coordinates, assumptions and calibration stand behind one claim?
Scope: Not applicable. The claim-level dependency graph is a declared acyclic bookkeeping graph; it does not identify and is not a causal graph.
Guarantee: claim_level_revalidation_relative_to_declared_dependencies_no_interval_or_calibration.
Bounds: max_nodes = 512, max_edges = 4096, max_overlap_checks = 250000.

revalidated interval

Claim: none

Question: Can an interval, posterior or coverage statement be reported for a revalidated or composed claim, or can a calibration record established for one claim, target population or composition be carried to another?
Scope: No graph.
Guarantee: no_interval_is_produced_by_verification.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

linear moment propagation

Claim: point_only

Question: What are the mean and variance of a linear transform, contrast, sum or weighted sum of uncertain inputs, and which dependence information must be supplied before they are formed?
Scope: Not applicable: summaries or aligned draws of declared scientific quantities.
Guarantee: exact_given_supplied_moments_point_only_no_interval.
Bounds: max_terms = 256.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

aligned nonlinear functional

Claim: point_only

Question: What is the expectation of a product or other nonlinear functional of two uncertain inputs, and when may their draws be paired?
Scope: Not applicable: aligned draws of declared scientific quantities.
Guarantee: aligned_draw_expectation_point_only_no_interval.
Bounds: max_draws = 100000.

known weight mixture

Claim: point_only

Question: What are the mean and variance of a finite mixture of uncertain components when the weights are known?
Scope: Not applicable.
Guarantee: exact_given_supplied_weights_and_moments_point_only_no_interval.
Bounds: none declared.

dependence envelope

Claim: none

Question: What values can a functional of two marginals take when nothing is known about their dependence, or when only a correlation range is declared?
Scope: Not applicable.
Guarantee: exact_structural_envelope_or_assumption_range_no_point_no_interval.
Bounds: max_marginal_atoms = 100000, max_lattice_checks = 4000000.

transported composition

Claim: none

Question: May a transported interval, or the source-sampling, target-sampling and transport-functional pieces of one transported claim, be composed with another claim?
Scope: As the transported claim's selection diagram; this route does not examine it.
Guarantee: composition_refused_without_aligned_law_or_covariance_point_only_otherwise.
Bounds: none declared.

composed interval

Claim: calibrated

Question: Can an interval or posterior statement be reported for a quantity composed from several uncertain inputs?
Scope: As the joint_law.* records.
Guarantee: linear_moment_composed_interval_measured_from_200_observations_posterior_and_other_operations_closed.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

scientific interface

Claim: none

Question: May the quantity one component produces feed the input another consumes, decided from the two declared contracts before any number is computed: do the variable, units, population, regime, support, measurement assumptions, available operations, uncertainty meaning, calibration and provenance agree, and is the evidence of the inputs known?
Scope: Not applicable. A composition is a declared directed graph of components; it is not a causal graph and the check does not identify.
Guarantee: decision_from_declarations_alone_no_numerical_execution_no_calibration_upgrade.
Bounds: max_ports = 256, max_components = 1024, max_connections = 8192, max_port_set_items = 1024.

cyclic composition

Claim: none

Question: Can components be composed when one component's output feeds back into another's input, directly or through a chain?
Scope: Not applicable. The graph is the declared connection graph of components.
Guarantee: cyclic_composition_refused_before_execution_with_the_cycle_path.
Bounds: max_components = 1024, max_connections = 8192.

calibration inheritance

Claim: none

Question: Can a consumer rely on measured calibration of an input, or a component declare its output's calibration inherited from an input, when the upstream producer does not hold a measured record?
Scope: Not applicable.
Guarantee: calibration_labels_carried_never_raised_inherited_calibration_resolved_or_refused.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: interface_contract_calibration_not_held).

portable specification

Claim: none

Question: Can a party that does not use Antecedent read and check the interface contract, evidence ledger and composition bundle artifacts: are their layouts, canonicalization and identity rules written down precisely enough that an independent implementation recomputes the same identities and rejects the same mutations?
Scope: Not applicable.
Guarantee: formats_specified_independently_consumed_identities_recomputed_no_numerical_claim.
Bounds: none declared.

finite markovian scm

Claim: point_only

Question: What would the same observed unit have had under another action, and what do worlds that share its exogenous history jointly do, in a declared finite Markovian SCM?
Scope: Static Markovian SCM over finite variables (at most 12, 4 levels, independent finite exogenous terms); semi-Markovian and nonparametric ADMG requests are refused.
Guarantee: exact_enumeration_point_only_conditional_on_declared_model.
Bounds: max_variables = 12, max_worlds = 8, max_units = 1024, max_joint_exogenous_states = 65536.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

temporal selective

Claim: point_only

Question: Can the fixed-population temporal counterfactual reuse its abduction when only the named action histories change?
Scope: Fully observed two-slice Markovian temporal DAG; no latent confounding.
Guarantee: shared_exogenous_history_counterfactual_linear_additive_noise_point_only_with_reuse.
Bounds: none declared.

transported row

Claim: none

Question: May a cross-world quantity be executed in a target population other than the one the mechanisms are declared for?
Scope: Not applicable (closed).
Guarantee: closed_with_named_prerequisites.
Bounds: none declared.

sampled and interval readings

Claim: calibrated

Question: May a cross-world value carry an interval, a posterior or a sampled abduction?
Scope: Finite Markovian SCM.
Guarantee: measured_interval_under_the_declared_model_from_60_units_posterior_and_sampled_abduction_closed.
Bounds: none declared.

law posterior protocol

Claim: none

Question: May a law, posterior or nested callback that Antecedent does not own be asked for its mean, draws, distribution function, quantile function or posterior update, each negotiated on its own, with the answer accepted only as a well-formed law of the declared quantity?
Scope: Not applicable. A callback is a declared provider object; the check does not identify.
Guarantee: declared_operations_negotiated_separately_output_accepted_only_as_a_well_formed_law_no_interval_no_native_trust.
Bounds: max_dependencies = 1024, max_inputs = 256, max_draws = 1000000, max_points = 65536.

replay policy

Claim: none

Question: When may a stored answer of a callback be reused for a request, when must it be recomputed, and when is the request refused: for deterministic pure, seeded stochastic, stateful and side-effecting callbacks, with dependencies changed, a provider replaced or hidden state changed?
Scope: Not applicable. A callback is a declared provider object; the check does not identify.
Guarantee: reuse_only_under_matching_identity_and_replay_contract_effects_never_invoked_outside_execution_or_retried_outside_policy.
Bounds: none declared.

portable result

Claim: none

Question: Can a callback result be carried to another process as a declaration and a portable output, consumed with every identity recomputed and every bound checked, resumed only by a supplied compatible provider, and otherwise recorded as an unavailable stage while the historical result stays inspectable?
Scope: Not applicable. A callback is a declared provider object; the check does not identify.
Guarantee: portable_declaration_and_output_identities_recomputed_resume_only_by_declared_provider_else_unavailable.
Bounds: none declared.

closed claims

Claim: none

Question: Can a callback supply an interval, a posterior or probability guarantee, native trust, an identification licence or a joint law?
Scope: Not applicable. A callback is a declared provider object; the check does not identify.
Guarantee: no_interval_no_posterior_guarantee_no_native_trust_no_identification_no_joint_law_from_a_callback.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed, joint_law_required).

consumers

Claim: none

Question: May an external provider's output be used as source evidence for a transport, a transported claim be handed to an external consumer, and a callback law be scored by the predictive assessment, each through the interface contract?
Scope: Not applicable. A callback is a declared provider object; the check does not identify.
Guarantee: connection_decided_from_declarations_before_invocation_diagnostic_callbacks_not_offered_to_assessment.
Bounds: none declared.

selective reuse

Claim: none

Question: When a discovery request is repeated with a different threshold, multiplicity rule, background knowledge, orientation budget, lag window, CI test or data, which conditional-independence statistics of the earlier run are reused, which stages are recomputed, what does the result carry about the equivalence class and exhausted bounds, and which downstream claims are still valid, recomputed or withdrawn?
Scope: Discovered structures keep their class: CPDAG, PAG, temporal DAG or temporal CPDAG. An equivalence class is never reported as one graph; unresolved edges and exhausted bounds are carried to every consumer.
Guarantee: algorithmic_match_with_full_run_bit_for_bit_no_recovery_or_error_control_claim.
Bounds: max_portable_receipts = 250000, max_structure_items = 65536.

selective reuse guarantees closed

Claim: none

Question: Can a per-dataset recovery or error-control statement, an interval for an estimate computed on the rows discovery used, score-search reuse, multi-population discovery or transport of a discovered structure be offered from selective discovery?
Scope: As discovery.selective_reuse.
Guarantee: closed_as_per_dataset_claims_faithfulness_and_sufficiency_untestable_recovery_published_only_as_a_measured_table.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

release boundary fixture

Claim: point_only

Question: Does one scenario that corrects a measured exposure, composes mechanisms, pools shared study evidence, decides from the resulting causal claim and revalidates it after new observations hold together through the public Rust and Python surfaces, with every unsupported conversion, withdrawal and dependence gap refused?
Scope: A finite two-variable structural model (exposure X, outcome Y) with declared tables; no graph search and no identification claim beyond the declared mechanisms.
Guarantee: point_only_composition_checked_against_closed_forms_by_an_independent_consumer_with_typed_refusals_for_every_mutation.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

covariate shift cohort effect

Claim: point_only

Question: What is the weighted mean treatment contrast of a retained fitted effect model over a declared target cohort in another population, when the only difference between the populations is declared to enter through the model's effect modifiers?
Scope: A selection diagram declared by the caller: one selection node whose children are a subset of the effect modifiers, which are exactly the model's features. The structure is checked, the invariance it states is not tested.
Guarantee: declared_selection_premise_weighted_cohort_mean_of_retained_contrast_point_only_box_overlap_checked.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

measurement model

Claim: point_only

Question: Can a measurement-error model established in a source population correct an indicator observed in a target population, and under which declared invariance of its error parameters?
Scope: Not applicable: the premise is an invariance of declared measurement-error parameters between named populations, not a graph claim.
Guarantee: declared_error_parameter_invariance_carried_model_correction_point_only.
Bounds: none declared.

fitted linear gaussian mechanism

Claim: point_only

Question: Which fitted linear-Gaussian mechanisms of a source population may be carried into a composition for a target population under a declared selection diagram, and what does the carried composition compute?
Scope: A declared acyclic causal graph over the composed variables and a selection diagram between a source and a target population; latent causes only as declared loadings.
Guarantee: diagram_checked_carry_of_fitted_linear_gaussian_mechanisms_first_order_source_fit_variance_labelled.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

synthesis pooling

Claim: point_only

Question: What is the fixed-effect pooled estimate of a native claim and claims transported into the same target population, with the transport error counted once per transport?
Scope: Not applicable: the combination consumes the evidence ledger's source declarations and the transport identities (populations, assumptions, selection diagram digest) of each claim.
Guarantee: gls_fixed_effect_pool_with_declared_transport_variance_counted_once_model_variance_no_interval.
Bounds: none declared.

synthesis borrowing

Claim: point_only

Question: What is the posterior mean and variance of a common effect when transported claims are borrowed with a declared between-claim heterogeneity prior and each transported variance is inflated by its transport error?
Scope: Not applicable: the combination consumes the evidence ledger's source declarations and the transport identities (populations, assumptions, selection diagram digest) of each claim.
Guarantee: hierarchical_borrowing_posterior_mean_and_variance_under_declared_heterogeneity_and_transport_variance_no_interval.
Bounds: none declared.

synthesis transported prior

Claim: point_only

Question: What is the posterior of a native likelihood when one transported claim is the prior, with its transport error carried once?
Scope: Not applicable: the combination consumes the evidence ledger's source declarations and the transport identities (populations, assumptions, selection diagram digest) of each claim.
Guarantee: normal_normal_posterior_of_a_native_likelihood_under_a_transported_prior_no_interval.
Bounds: none declared.

surrogate bridge premise

Claim: point_only

Question: Can a trial-level surrogate-to-terminal mapping validated in one population carry a surrogate effect into a claim about another population, and which selection-diagram premise licenses it?
Scope: A selection diagram declared between the validation population and the claim's population over the terminal outcome and the declared relation modifiers.
Guarantee: diagram_checked_invariance_premise_trial_level_mapping_point_mean_inside_validated_range.
Bounds: none declared.

policy evaluation

Claim: point_only

Question: What is the value of a supplied contingent policy in a target population when the only difference from the model's population is the initial-state law, under a declared selection diagram over every model mechanism?
Scope: A finite sequential model with a declared selection diagram assigning invariant or shifted to the initial-state law and to every transition, observation and utility mechanism; only the initial-state law may be shifted.
Guarantee: declared_selection_diagram_only_initial_state_law_shifted_exact_finite_enumeration_point_only.
Bounds: none declared.

period shift policy

Claim: point_only

Question: What is the value of a supplied contingent policy in a later period when mechanisms are declared time-invariant over a fixed horizon and only the initial-state law changes?
Scope: A finite sequential model with a declared per-mechanism time-invariance status and one fixed horizon equal across model, declaration and target period.
Guarantee: declared_time_invariance_fixed_horizon_identical_alphabets_exact_finite_enumeration_point_only.
Bounds: none declared.

adaptive trial source

Claim: point_only

Question: What are the arm means of an adaptively assigned trial standardized to a target population's stratum law, using the recorded assignment probabilities, under a declared selection diagram?
Scope: Finite strata with a declared selection diagram marking the stratum law shifted and the outcome mechanism given stratum and arm invariant.
Guarantee: recorded_probability_hajek_standardization_to_target_stratum_law_point_only_uncalibrated_no_variance.
Bounds: none declared.

cpdag class envelope

Claim: structural_envelope

Question: Is a transport query identified in every member of the Markov equivalence class a discovery run returned, and by which expressions?
Scope: A static CPDAG with causal sufficiency, at most 64 undirected edges and at most the declared member bound of enumerated members; a PAG is refused.
Guarantee: structural_envelope_over_enumerated_equivalence_class_members_no_point_no_interval.
Bounds: none declared.
Not an interval: the structural envelope is never a confidence interval.

per population

Claim: none

Question: What structure does one discovery contract return on each of several populations searched separately, and where do the structures differ?
Scope: The contract's declared structure class per population (static); no cross-population graph is estimated.
Guarantee: independent_per_population_runs_of_one_contract_no_pooling_differences_descriptive_only.
Bounds: none declared.

sample split analysis

Claim: none

Question: Can discovery and estimation use disjoint rows of one static table with a recorded, re-derivable split, so an analysis interval is at least conditional on the selected graph?
Scope: A static tabular float64 table without weights or masks; dependent rows are not covered.
Guarantee: disjoint_row_split_with_rederivable_receipt_interval_conditional_on_selected_graph_no_coverage_claim.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

sample split effect

Claim: point_only

Question: After discovery on one half of a table, what is the linear effect of a treatment on an outcome estimated by adjustment on the disjoint other half, and over which measured family of generating processes does its interval cover the true effect?
Scope: A static float64 table of five to seven searched variables whose generating DAG is linear-Gaussian, causally sufficient and faithful, with a fully orientable equivalence class.
Guarantee: whole_method_coverage_of_the_true_effect_measured_over_a_declared_family_never_a_per_dataset_recovery_statement.
Bounds: min_variables = 5, max_variables = 7.
Interval: withheld. The interval route is closed (cell_not_licensed) until its coverage records are measured; the result is the point only.

provider source admission

Claim: none

Question: When may the output of an external provider (a law, or a parameter or causal-functional posterior) be admitted as source evidence for a transported claim, and with what standing?
Scope: Not applicable: the admission checks declarations against a transport identity; no graph is estimated.
Guarantee: declared_population_regime_assumptions_trust_checked_source_evidence_never_native_target_data_never_measured.
Bounds: none declared.

handoff artifact

Claim: none

Question: Can a transported claim be handed to a consumer or a declared callback input with its transport identity retained and its standing stated as transported, and can that handoff be sealed and re-derived by a fresh consumer?
Scope: Not applicable: the transport identity carries the selection diagram digest and assumptions; no graph is estimated here.
Guarantee: transported_standing_retained_consumer_acceptance_required_sealed_form_rederived_never_trusted.
Bounds: none declared.

procedure law

Claim: none

Question: What is the exact law of the terminal state of a fully declared adaptive procedure under a hypothesized truth or a Beta prior predictive?
Scope: Not applicable: the result is a property of the declared collection procedure (assignment rule, enrolment schedule, outcome delay and loss, exposure failure, stopping rule, final selection) under a hypothesized truth or a prior predictive.
Guarantee: exact_finite_law_of_the_declared_procedure_under_a_hypothesized_truth_or_prior_predictive_no_interval_or_coverage_license.
Bounds: max_arms = 4, max_contexts = 4, max_cells = 8, max_looks = 24, max_units = 400, max_outcome_lag = 4.

procedure bayesian martingale gap

Claim: point_only

Question: Does a declared Beta prior's posterior mean stay a martingale of the whole adaptive procedure, and by how much does it fail when delay or loss depends on outcome values?
Scope: Not applicable: the result is a property of the declared collection procedure (assignment rule, enrolment schedule, outcome delay and loss, exposure failure, stopping rule, final selection) under a hypothesized truth or a prior predictive.
Guarantee: exact_prior_predictive_posterior_mean_gap_and_closed_form_beta_report_derived_unmeasured_no_credible_interval_coverage.
Bounds: none declared.

procedure decision report

Claim: point_only

Question: What are the stopping-time law, expected sample size, decision law, probability of a wrong decision and expected regret of a declared expected-loss stopping procedure under a hypothesized truth?
Scope: Not applicable: the result is a property of the declared collection procedure (assignment rule, enrolment schedule, outcome delay and loss, exposure failure, stopping rule, final selection) under a hypothesized truth or a prior predictive.
Guarantee: exact_stopping_time_and_decision_law_wrong_decision_probability_and_expected_regret_under_a_hypothesized_truth_derived_unmeasured.
Bounds: none declared.

procedure anytime exclusion

Claim: point_only

Question: Do time-uniform confidence-sequence bounds at the data-dependent stopping time of a declared adaptive procedure exclude the truth with probability at most alpha, and what is that probability exactly?
Scope: Not applicable: the result is a property of the declared collection procedure (assignment rule, enrolment schedule, outcome delay and loss, exposure failure, stopping rule, final selection) under a hypothesized truth or a prior predictive.
Guarantee: exact_whole_procedure_exclusion_probability_of_the_time_uniform_bounds_at_a_declared_truth_derived_unmeasured.
Bounds: none declared.

procedure ordered p value

Claim: point_only

Question: What are the stagewise-ordered tail probabilities of the observed terminal state of a fixed-randomization sequential procedure under the exact whole-procedure law, and the normal-model stagewise median-unbiased inference of a stopped statistic?
Scope: Not applicable: the result is a property of the declared collection procedure (assignment rule, enrolment schedule, outcome delay and loss, exposure failure, stopping rule, final selection) under a hypothesized truth or a prior predictive.
Guarantee: exact_whole_procedure_ordered_tail_probabilities_and_grid_inversion_set_conditional_on_the_declared_procedure_derived_unmeasured.
Bounds: max_inversion_laws = 4096.

procedure weighting simulation

Claim: point_only

Question: What is the sampling law of an adaptively weighted contrast estimate under the stopping time, delay, loss and policy selection of a declared procedure, by whole-procedure simulation?
Scope: Not applicable: the result is a property of the declared collection procedure (assignment rule, enrolment schedule, outcome delay and loss, exposure failure, stopping rule, final selection) under a hypothesized truth or a prior predictive.
Guarantee: seeded_whole_procedure_simulation_of_the_weighted_estimator_law_with_monte_carlo_error_no_coverage.
Bounds: max_replications = 1000000.

derived numerical interval

Claim: none

Question: What numerical (Monte Carlo) interval does a candidate's expected utility, a decision-minus-candidate difference or an event probability carry at the effective sample size of the supplied draws?
Scope: Not applicable: the statement is about draws of utilities of a finite candidate set whose law is the supplied one.
Guarantee: monte_carlo_wald_and_wilson_intervals_at_the_effective_sample_size_derived_unmeasured_never_summed_with_source_spread.
Bounds: none declared.

optimality confidence set

Claim: none

Question: Which retained candidates cannot be excluded as the best, at a stated level, from paired or independent difference errors?
Scope: Not applicable: the statement is about draws of utilities of a finite candidate set whose law is the supplied one.
Guarantee: multiple_comparisons_with_the_best_confidence_set_for_the_best_retained_candidate_derived_unmeasured.
Bounds: none declared.

selection corrected interval

Claim: none

Question: What interval for the value of the selected candidate accounts for having selected the best of the n considered candidates, beside the held-out interval and the winner's-curse bound?
Scope: Not applicable: the statement is about draws of utilities of a finite candidate set whose law is the supplied one.
Guarantee: bonferroni_selection_corrected_interval_with_held_out_interval_and_winners_curse_bound_derived_unmeasured.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

producer carrier

Claim: none

Question: Can the outputs of the cohort decision, the sequential rollout and the off-policy evaluation reach the reliability diagnostics, and what must they retain to do so?
Scope: Not applicable: the carriers consume utility draws and their declared sharing.
Guarantee: per_draw_utilities_from_one_declared_law_reach_the_carrier_point_only_outputs_refused_ranking_only_for_structural_noise.
Bounds: none declared.

composed spread linear

Claim: none

Question: What is the spread (variance, standard error) of a linear composition of component estimates, labelled uncalibrated, and which dependence licenses it?
Scope: Not applicable: the dependence between composed claims is a declared joint-law input, not a graph.
Guarantee: delta_method_variance_of_a_linear_composition_under_a_licensed_dependence_labelled_uncalibrated_spread_no_interval.
Bounds: none declared.

composed spread aligned functional

Claim: none

Question: What is the spread of a nonlinear composed functional (product, ratio, min, last stage of a chain) of aligned component draws?
Scope: Not applicable: the dependence between composed claims is a declared joint-law input, not a graph.
Guarantee: aligned_row_spread_of_a_nonlinear_functional_labelled_uncalibrated_interval_candidates_unmeasured.
Bounds: none declared.

composed exceedance

Claim: none

Question: What is the probability that a composed quantity exceeds a cutoff, from a known-weight mixture or aligned draws, and what does the Monte Carlo band license?
Scope: Not applicable: the dependence between composed claims is a declared joint-law input, not a graph.
Guarantee: weighted_exceedance_and_monte_carlo_banded_verdict_of_a_composed_quantity_uncalibrated_posterior_reading_closed.
Bounds: none declared.

assumption range

Claim: assumption_range

Question: What range of a transported quantity follows from a bounded declared violation of its transport premise, and at which shift does a decision on it change?
Scope: Not applicable: the premise violation is a declared bounded shift of the transported quantity, for each of 13 declared transport routes with a scalar value.
Guarantee: closed_form_range_under_a_declared_bounded_premise_violation_assumption_range_no_probability_no_sampling_composition.
Bounds: none declared.
Not an interval: the assumption range is never a confidence interval.

revalidation

Claim: none

Question: Which transported claims remain valid, must be recomputed, or are withdrawn when a source dataset, the selection diagram, the target population or a calibration record changes?
Scope: A declared claim graph: target, selection diagram, sources, transported claims and decisions, with claim coordinates hashed from target and quantity.
Guarantee: three_state_classification_of_transported_claims_by_graph_re_derivation_no_new_license_calibration_never_inherited_across_targets.
Bounds: none declared.

transported source

Claim: point_only

Question: Can a transported point or law, with its transport receipt, be the source of a decision, and under which acceptance, support and uncertainty rules?
Scope: Not applicable: the source populations and transport identities are carried by the receipt; no graph is estimated by the decision.
Guarantee: transported_value_with_receipt_decides_by_affine_expected_utility_or_a_supplied_law_never_inflated_labels_no_interval.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

transported decision artifact

Claim: point_only

Question: Can a transported point decision be exported and consumed so that a fresh consumer re-evaluates it rather than trusting stored numbers?
Scope: Not applicable.
Guarantee: point_only_export_consumer_re_evaluates_and_compares_identity_and_action_values_law_not_exportable.
Bounds: none declared.

composed forward source

Claim: point_only

Question: Can a composed mechanism model (linear-Gaussian or finite) supply the per-action forward claim of a finite-action inverse query, and under which conditions?
Scope: Not applicable: the forward source is a value of the action response; its causal standing is carried by the origin.
Guarantee: replayed_composed_mean_or_exact_finite_law_per_action_carried_mechanisms_need_receipt_and_acceptance_point_only.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

cohort forward source

Claim: point_only

Question: Can the weighted cohort-mean contrast of a retained fitted effect model be the forward claim of a finite-action inverse query?
Scope: Not applicable: the forward source is a value of the action response; its causal standing is carried by the origin.
Guarantee: cohort_mean_contrast_as_a_point_forward_source_shifted_cohort_is_transported_and_needs_acceptance.
Bounds: none declared.

transported forward source

Claim: point_only

Question: Can a transported response (a declared point or law, a transported policy value, a transported synthesis) be the forward claim of a finite-action inverse query?
Scope: Not applicable: the forward source is a value of the action response; its causal standing is carried by the origin.
Guarantee: transported_response_as_a_forward_source_under_explicit_acceptance_receipt_bound_into_the_origin_point_only.
Bounds: none declared.

observational scenarios

Claim: point_only

Question: Given an observational joint law of covariates and an outcome, how likely is each scenario of a covariate event given an outcome event, and the converse, as an associational description?
Scope: Not applicable: associational; no graph and no intervention is used.
Guarantee: weighted_bayes_inversion_of_a_supplied_aligned_observational_law_associational_not_an_intervention_effect.
Bounds: none declared.

discretized linear gaussian scm

Claim: point_only

Question: Can a composed linear-Gaussian mechanism model be discretized into an exact finite SCM, with a computed bound on the error of a mean contrast, so finite-law routes can consume it?
Scope: A Markovian acyclic linear-Gaussian composition over the real line (no latent loadings), 2 to 64 levels per variable and at most 2^20 joint states.
Guarantee: exact_finite_enumeration_of_the_discretized_model_with_a_computed_mean_contrast_error_bound_point_only.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

discretized scm artifact

Claim: point_only

Question: Can a discretized composition be exported and consumed so that a fresh consumer re-derives it from the carried premises rather than trusting stored numbers?
Scope: As for the discretized model.
Guarantee: consumer_re_derives_identity_and_every_claimed_number_bit_for_bit_point_only.
Bounds: none declared.

paradigm parity

Claim: none

Question: For each route added with the transport and uncertainty work, which inferential paradigms (frequentist, Bayesian) are supported and which are a typed closed row with a reason?
Scope: Not applicable.
Guarantee: registry_decision_every_route_has_a_row_and_every_missing_paradigm_is_a_typed_closed_row.
Bounds: none declared.

whole path uncertainty

Claim: none

Question: What is the uncertainty of a finite Markovian cross-world result over its units, split into structural, exogenous-posterior, sampling and numerical components?
Scope: A finite Markovian structural causal model with declared mechanisms and enumerated exogenous states (the cross-world execution row).
Guarantee: enumerated_exogenous_posterior_spread_with_declared_unit_sampling_components_never_added_uncalibrated_no_interval.
Bounds: none declared.

component adapters

Claim: none

Question: Can a declared mechanism, scientific bridge or measurement model take part in the one interface-contract composition check, and what must it state beyond its own declaration?
Scope: Not applicable: the composition is a declared set of components and connections; a cycle is refused with its path.
Guarantee: declared_component_ports_checked_by_the_one_shared_check_no_inferred_facts_domain_licensing_stays_with_the_component.
Bounds: none declared.

ledger overlap

Claim: none

Question: Is a DisjointEvidence requirement between composed components met, decided from the evidence ledger's observation rows rather than from source identities?
Scope: Not applicable.
Guarantee: disjoint_evidence_decided_from_ledger_rows_unknown_overlap_is_not_disjoint_ledger_free_check_refuses_rather_than_skips.
Bounds: none declared.

ledger independence

Claim: none

Question: Is a declared independence between mechanisms' parameter estimates decided from the evidence ledger's rows rather than from evidence strings?
Scope: A declared composition with per-mechanism evidence ids read as ledger source identifiers.
Guarantee: independence_decided_from_ledger_rows_unknown_overlap_is_not_independent_string_equality_only_defence_in_depth.
Bounds: none declared.

ledger overlap and row pairing

Claim: none

Question: Are two propagation inputs independent or pairable by row according to the ledger and their real alignment identities, rather than by strings or draw counts?
Scope: Not applicable.
Guarantee: independence_from_ledger_rows_and_row_pairing_from_real_alignment_identities_no_index_pairing.
Bounds: none declared.

composed spread transported

Claim: none

Question: What is the spread of a transported composition of source sampling, target sampling and the transport functional, and which dependence license it?
Scope: Not applicable: the dependence between the pieces is a declared covariance or an aligned joint law.
Guarantee: first_order_spread_of_a_transported_composition_with_all_three_pieces_and_a_licensed_dependence_uncalibrated_no_interval.
Bounds: none declared.

composed interval check

Claim: none

Question: Does a measured coverage record at the composed coordinate and construction license a composed interval, and what is returned without one?
Scope: Not applicable.
Guarantee: composed_interval_opens_only_with_a_measured_record_at_its_own_coordinate_and_construction_closed_otherwise.
Bounds: none declared.
Interval: withheld. The sampling-uncertainty interval is not offered (its route is closed: cell_not_licensed).

procedure repeated measures

Claim: none

Question: What is the law and the derived uncertainty of a contrast estimated from repeated or clustered Gaussian measurements of units under a declared adaptive or fixed assignment, with the within-unit dependence declared?
Scope: Not applicable: a property of the declared collection procedure and within-unit dependence structure.
Guarantee: exact_closed_forms_and_a_seeded_simulation_of_the_declared_repeated_measures_procedure_derived_unmeasured_no_coverage.
Bounds: none declared.

Explicit refusals

No root query is without a licensed cell.

A result is available only at its declared scientific and support coordinate. See the support matrix for exact coordinates and reasons.

Analysis support changes since 2.3

Changes to generic analysis support since 2.3.0: 196 added coordinates, 0 removed coordinates.

Added analysis coordinates

AnomalyAttribution × TemporalDag × explicit × Frequentist × validation none
AnomalyAttribution × TemporalDag × explicit × Bayesian × validation none
AverageDerivative × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
AverageDerivative × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
AverageEffect × TemporalAdmg × explicit / accepted / graph_posterior × Frequentist × validation none / cheap / full
AverageEffect × TemporalAdmg × explicit / accepted / graph_posterior × Bayesian × validation none / cheap / full
ChangeAttribution × TemporalDag × explicit × Frequentist × validation none
ChangeAttribution × TemporalDag × explicit × Bayesian × validation none
Counterfactual × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
Counterfactual × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
DirectionalDerivative × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
DirectionalDerivative × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
Elasticity × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
Elasticity × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
InterventionalDistribution × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
InterventionalDistribution × TemporalAdmg × explicit / accepted × Frequentist × validation none
InterventionalDistribution × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
InterventionalDistribution × TemporalAdmg × explicit / accepted × Bayesian × validation none
InterventionResponse × TemporalAdmg × explicit / accepted / graph_posterior × Frequentist × validation none / cheap / full
InterventionResponse × TemporalAdmg × explicit / accepted / graph_posterior × Bayesian × validation none / cheap / full
NestedCounterfactualEffect × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
NestedCounterfactualEffect × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
PathSpecificEffect × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
PathSpecificEffect × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
PointDerivative × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
PointDerivative × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
ResponseCurve × TemporalAdmg × explicit / accepted / graph_posterior × Frequentist × validation none
ResponseCurve × TemporalAdmg × explicit / accepted / graph_posterior × Bayesian × validation none
ResponseJacobian × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
ResponseJacobian × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
SemiElasticity × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
SemiElasticity × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
TemporalConditionalPulse × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
TemporalConditionalPulse × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
TemporalConditionalSustained × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Frequentist × validation none
TemporalConditionalSustained × TemporalCpdag / TemporalPag × explicit / accepted / graph_posterior × Bayesian × validation none
TransportQuery × TemporalAdmg × explicit × Frequentist × validation none
TransportQuery × TemporalAdmg × explicit × Bayesian × validation none

See the support matrix for the complete current inventory and interval standing.

What's Changed

Release 2.4.0 by @nulldiver in #64

Full Changelog: v2.3.0...v2.4.0