{"actors":["Athletes who generate training, recovery, symptom, and perceived-exertion observations","Coaches who prescribe sessions and make participation decisions","Strength and conditioning staff who record completed work and adjust plans","Sport scientists who define workload and recovery projections","Medical and physiotherapy staff who contribute restricted availability events","Data stewards who govern identity, access, retention, and correction"],"affected_objective":"Produce an explainable, decision-time view of each athlete's recent training exposure, recovery evidence, and data completeness for daily training planning without treating that view as a diagnosis or automatic participation decision.","arm":"COMMON_P1","authority_safety":{"authorized_first_step":"With athlete consent and data-governance approval, build a read-only prototype over a copied, de-identified four-week dataset; it may generate comparison reports but cannot write to operational records or recommendations.","decision_authority":"The head coach retains training-plan authority, qualified medical staff retain diagnosis and return-to-play authority, athletes retain applicable consent and contestation rights, and the designated data steward controls event admission, identity linkage, access, and redaction.","excluded_actions":["Automatically clearing, restricting, selecting, penalizing, or ranking athletes","Inferring injury, intent, effort, mental state, or causality from temporal sequence alone","Collecting continuous location, sleep, or off-duty behavior beyond the approved purpose","Exposing medical or athlete-reported details to unauthorized coaching or performance personnel","Silently merging uncertain athlete identities, overwriting disputed records, or editing derived projections as independent truth","Using the prototype for live training or return-to-play decisions"],"halt_rollback":"Stop the prototype if identity collisions, unauthorized sensitive-field exposure, untraceable source records, or replay disagreement on a declared deterministic projection occurs. Revoke prototype access, discard derived views and copied working data under the approved deletion procedure, preserve only the permitted incident audit, and leave source systems unchanged."},"baseline":"Separate coaching plans, wearable summaries, session-RPE entries, wellness forms, and medical availability tables feed an athlete-day dashboard whose current values can be overwritten or manually patched. Source timestamps and corrections are inconsistently retained, so staff cannot reliably reproduce what the dashboard showed at an earlier decision time or distinguish a missed session from missing telemetry.","candidate_id":"event_log_centered_modeling__sport_science__COMMON_P1","causal_chain":["A prescribed session, completed training bout, sensor upload, athlete report, restricted clinical availability assertion, and later correction arrive from different actors and clocks.","Updating one athlete-day record collapses occurrence time, reporting time, provenance, uncertainty, and correction history into a mutable current value.","Late uploads, duplicate imports, revised session classifications, and disputed identity links can therefore change or multiply workload and recovery summaries without an explicit lineage.","The intervention admits each bounded assertion once under a stable identity, retains both event and record time, and appends linked corrections instead of overwriting accepted history.","Pinned projectors replay eligible events under declared ordering, access, and uncertainty rules to create coaching, sport-science, and restricted medical views.","Displayed frontiers, source coverage, unresolved conflicts, and rebuild diffs reveal when a view is lagging, incomplete, or inconsistent with the canonical history.","Staff can then inspect the contributing evidence and exercise their existing authority using an explainable decision-time view rather than an opaque mutable snapshot."],"cell_id":"event_log_centered_modeling__sport_science","consequence":"A dashboard may present an apparently current workload or recovery state that cannot be traced to the exact prescribed and completed sessions, reports, corrections, and missing sources that produced it. This can support inconsistent training adjustments, obscure whether a value was known when a decision was made, and make retrospective analysis sensitive to duplicate or late data.","diversity_from_prior_proposals":"Not assessed because runtime isolation prohibits inspecting prior proposals; this candidate is derived only from the supplied archetype and sport-science domain card.","experiment_id":"eoa_inverse_innovation_exp13_second_slot_policy60_20260806","intervention":"Create a governed athlete-exposure event ledger in which canonical records represent semantically bounded assertions such as session prescribed, training bout completed, session classification reported, device summary received, perceived exertion reported, wellness response recorded, availability status asserted, and record corrected or adjudicated. Each record carries a stable event/assertion identifier, athlete reference, participant roles, event time, record time, training context, transformation or evidentiary delta, provenance, uncertainty, validity status, and access class. Raw high-frequency telemetry remains outside the semantic ledger unless a declared projection requires it; admitted device summaries link to retained source provenance. Duplicate submissions are idempotent, uncertain identity matches remain unresolved, and corrections append supersession or retraction links. Versioned projectors derive athlete trajectories, session-load summaries, decision-time views, and completeness indicators; each view exposes its projector version, frontier, source coverage, conflicts, and late-event status and can be rebuilt and compared with the live projection.","mechanism_mapping":[{"counterfactual_removal":"Without stable append-only acceptance, retries and corrections can multiply or silently replace training evidence, so decision-time reconstruction fails.","mechanism_slug":"append_only_event_store","role":"Makes accepted sport-science event assertions and their correction lineage the governed historical substrate."},{"counterfactual_removal":"Without separate occurrence and record times, a late sensor upload or retrospective exertion report appears to have been available during an earlier coaching decision.","mechanism_slug":"bitemporal_event_register","role":"Separates when training or recovery evidence pertains from when it was submitted, imported, corrected, or adjudicated."},{"counterfactual_removal":"Without stable identities and idempotent application, overlapping device imports or form retries can double-count the same bout or report.","mechanism_slug":"event_replay_deduplication","role":"Prevents duplicate ingestion and replay from multiplying workload contributions."},{"counterfactual_removal":"Without versioned projections, workload and recovery values remain independently mutable summaries whose inputs and interpretation cannot be reproduced.","mechanism_slug":"event_sourced_projection","role":"Derives audience-specific athlete trajectories and current views from eligible canonical events under declared rules."},{"counterfactual_removal":"Without rebuild-and-diff, manual patches or non-deterministic projector behavior can persist undetected in operational dashboards.","mechanism_slug":"projection_rebuild_and_diff","role":"Recomputes a projection from the ledger and compares it with the live result to expose drift."},{"counterfactual_removal":"Without visible frontiers and coverage, users may interpret a view missing a device partition, questionnaire, or correction as complete and current.","mechanism_slug":"projection_frontier_dashboard","role":"Shows processing lag, contributing-source coverage, unresolved conflicts, and the latest fully represented event boundary."}],"nearest_rivals":["A conventional athlete-management system with mutable athlete-day records and an audit trail: it may show edits, but the audit trail need not contain sufficient typed transformations, ordering rules, or projector versions to rebuild every view.","A time-series data warehouse joining wearable, questionnaire, and session tables: it supports longitudinal queries, but source tables and materialized summaries may remain independently authoritative and corrections may not preserve decision-time knowledge states.","A training diary or athlete timeline: it presents chronological entries, but may omit participant roles, dual time, deduplication, explicit uncertainty, correction lineage, replay contracts, and projection frontiers.","A versioned daily readiness record: it retains successive snapshots, but snapshots rather than bounded multi-actor events remain primary and cannot necessarily explain how several views arose from shared history."],"negative_tests":{"intervention_falsifier":"The intervention is falsified for this use if, after event definitions and source mappings are frozen, a pinned replay cannot reproduce the reference session-load and decision-time views from admitted events within predeclared exact-match rules, or if required current values depend on unrecorded mutable state that cannot be represented without unacceptable capture.","problem_falsifier":"The problem is falsified if the existing systems can already reproduce any sampled athlete's current and prior decision-time views from immutable, uniquely identified, provenance-bearing records; distinguish event time from record time; trace corrections and duplicates; and display source completeness without manual reconstruction.","risks":["Event granularity may create excessive capture burden or retain surveillance-like detail.","Missing devices, informal training, non-reporting, and excluded athletes may be mistaken for absence of activity.","Identity resolution errors may attach sensitive or performance-relevant events to the wrong athlete.","Coach-entered and device-derived sources may receive unjustified evidentiary weight over athlete reports.","A workload trajectory may be misread as causal evidence of injury, motivation, or responsibility.","Projection lag or incorrect ordering may make an incomplete view appear actionable.","Long-lived event history may expose health, employment, or behavioral information beyond its stated purpose.","Schema choices may exclude adaptive sport, nonstandard training, or context-specific exertion and recovery experiences.","Append-only correction may conflict with legitimate deletion or redaction duties unless propagation is tested."] ,"strongest_counterevidence":"Direct current-state measurement may be sufficient for daily planning, while historical replay could add capture, governance, and privacy costs without changing any authorized decision. Continuous physiological signals may also resist defensible event boundaries, making a discrete semantic ledger a lossy representation."},"next_evidence_step":"Run a bounded retrospective replay on six consented, de-identified athlete records covering four weeks and three sources: prescribed sessions, completed-session/device summaries, and session-RPE reports. Before examining results, freeze event definitions, deduplication keys, ordering rules, one session-load projector, one decision-time projector, and exact comparison criteria. Compare reconstructed outputs with archived source-derived reference values; count unexplained mismatches, duplicate suppressions, late events, missing-source periods, and replay-versus-live diffs. Inspect no free-text medical notes, make no athlete-level performance judgment, and stop after the predefined sample.","observable_state":"For a selected athlete and decision timestamp, staff can see the current workload or recovery summary but cannot consistently enumerate which session occurrences, imports, subjective reports, exclusions, corrections, and missing sources contributed; rerunning an import or accepting a late record can alter the displayed trajectory without an explicit frontier or preserved prior view.","prior_art_status":"UNSEARCHED","problem":"Daily training decisions rely on athlete-level workload and recovery summaries assembled asynchronously from planned sessions, completed work, wearable-derived summaries, perceived-exertion reports, wellness responses, and restricted availability information. When these inputs update mutable athlete-day records, late submissions, duplicate imports, reclassified sessions, and corrections can silently rewrite the trajectory. The organization cannot reliably reconstruct what evidence was available at a past decision time, why two views disagree, or whether an apparent zero represents no activity or an unobserved source.","proposal_index":1,"remaining_contrastive_claim":"For sport-science monitoring that must reproduce prior decision-time views, the defining difference is to make bounded, provenance-bearing training and recovery assertions canonical and make athlete-day summaries discardable, versioned projections with visible frontiers. Merely adding timestamps, an audit trail, or a longitudinal dashboard does not satisfy the claim unless the accepted history alone can rebuild the views under declared ordering and correction rules.","revision_record":{"claim_changes":[],"conceptual_changes":[],"evidence_changes":[],"operational_changes":[],"parent_version":null,"progress_targets_addressed":["Concrete sport-science problem and affected decision objective specified","Event boundaries, actors, observable state, and consequence made explicit","Structural and mechanism mappings preserve event-first causal structure","Authority, privacy safeguards, falsifiers, risks, and rollback bounded","First evidence step limited to retrospective, de-identified replay testing"]},"schema_version":1,"structural_mapping":[{"archetype_element":"Canonical event assertion","domain_realization":"A bounded sport-science assertion such as session prescribed, bout completed, device summary received, exertion reported, availability asserted, or prior assertion corrected, with athlete and contributor roles."},{"archetype_element":"Stable event identity and participant reference","domain_realization":"Source-scoped event identifiers and persistent pseudonymous athlete references support deduplication while uncertain device-to-athlete or session matches remain unresolved rather than forced."},{"archetype_element":"Event time and record time","domain_realization":"The time or interval of the training or recovery observation is stored separately from submission, import, validation, correction, and adjudication time."},{"archetype_element":"Transformation delta, provenance, and status","domain_realization":"Each event states what evidence or projected state contribution changed, which person or system supplied it, how it was derived, its uncertainty, access class, and whether it is reported, inferred, disputed, accepted, superseded, or retracted."},{"archetype_element":"Ordering and concurrency semantics","domain_realization":"Within-session sequence, source-local order, event-time intervals, causal claims, and computational tie-breaks are distinct; concurrent reports are not falsely presented as one causal chronology."},{"archetype_element":"Append-only correction and late-event policy","domain_realization":"Reclassified sessions, erroneous reports, and corrected device summaries link to prior assertions; material late events trigger scoped projection rebuilds while frozen decision-time views remain reproducible."},{"archetype_element":"Versioned projection definition","domain_realization":"Named projectors declare eligible event types, role interpretation, workload calculation, uncertainty handling, ordering, access filter, reference data, and output contract for each audience."},{"archetype_element":"Projection frontier and completeness statement","domain_realization":"Each view displays the latest fully processed boundary, lagging sources, unprocessed corrections, known device outages, non-reporting, and event classes outside coverage."},{"archetype_element":"Replay and reconciliation","domain_realization":"A pinned projector rebuilds athlete trajectories from accepted events and compares hashes, totals, and selected invariants with the live view; discrepancies lead to repair or rebuild rather than direct patching."},{"archetype_element":"Retention, redaction, and differential access","domain_realization":"Purpose-limited fields, role-specific projections, retention horizons, athlete contestation, and tested deletion propagation constrain exposure across the ledger, projections, indexes, snapshots, and exports."}],"title":"Decision-Time Athlete Exposure Ledger","version":0}