{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp06_four_proposal_generalization60_20260803","cell_id":"bounded_rivalry_governance__futurism_foresight","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_id":"brg-foresight-sealed-forecast-promotion-league","proposal_index":4,"version":0,"title":"Sealed-Forecast Promotion League","problem":"A strategic-foresight office has forty trained forecasters but only eight funded seats on the standing council that supplies probability estimates to decision teams. Selection through an open voluntary leaderboard can reward choosing easy questions, abstaining from difficult ones, copying visible consensus, updating only after others reveal information, using unequal research resources, coordinating forecasts, or influencing how questions resolve. A high rank can therefore reflect control of exposure and contest conditions rather than consistently informative forecasting.","actors":["Strategic-foresight office","Eligible staff forecasters","External subject-matter forecasters admitted under the same rules","Eight-member standing forecast council","Independent question-writing team","Independent resolution committee","Tournament administrator","Forecast audit and integrity staff","Independent appeal reviewer","Decision teams receiving aggregate forecasts","People or organizations potentially affected by forecast publication"],"observable_state":"The office can observe question assignments, completion and abstention rates, forecast timestamps, update timing, research-credit use, pairwise forecast similarity, communications among contestants, conflicts and outcome-influence opportunities, scores by question difficulty and horizon, audit corrections, resolution disputes, promotion and relegation patterns, seat concentration, aggregate forecast use, confidentiality incidents, and whether entrants repeatedly avoid or underperform on particular question classes.","consequence":"Council seats may go to contestants who manage question exposure, copy information, exploit resolution rules, or outspend rivals rather than those whose independent estimates add useful information across strategic uncertainties. The leaderboard can encourage herding, excessive research effort, concealment, and attempts to affect either outcomes or adjudication, while repeated winners acquire privileged access that helps preserve their rank.","affected_objective":"Use rivalrous forecasting to identify a rotating, plural council of forecasters whose sealed estimates remain comparable and decision-relevant across assigned uncertainties, without turning ranking, resource expenditure, collusion, or outcome influence into the easiest route to a seat.","intervention":"Create a seasonal promotion-and-relegation forecasting league. The scarce prize is four of the eight funded council seats each season; the other four provide staggered continuity and become contestable in the following season. Before play, an independent administrator freezes eligibility, question strata, assignment procedures, forecast windows, update limits, scoring, tie-breaks, audits, fouls, resolution rules, appeals, and promotion thresholds. Question writers and resolvers cannot compete. Eligible contestants receive balanced question blocks by lottery across topic, horizon, uncertainty, and anticipated resolution difficulty; they must answer the complete assigned block, subject only to disclosed conflict or outcome-influence recusals. Probabilities remain sealed from rivals until the relevant window closes. Contestants receive equal research-tool credits and the same number of scheduled updates. Scores use a precommitted proper scoring rule, are normalized within comparable question strata, and require a minimum completed-question count. Council promotion depends on a rolling two-season record rather than one question or one short streak. Leading accounts are audited for identity, timestamps, data access, resource use, recusals, and forecast independence before seats are awarded. Statistical screens examine suspiciously matched probabilities, synchronized updates, reciprocal information sharing, and stable seat rotation, but only refer anomalies for investigation. Published graduated fouls cover sharing sealed entries, proxy accounts, privileged-data misuse, evading resource caps, manipulating outcomes, lobbying resolvers, and harassing rivals. Contestants can challenge assignment, scoring, audit, or resolution errors through a time-boxed independent appeal. Only aggregate forecasts are supplied to decision teams; individual standings are delayed and restricted to the league unless publication is separately authorized. Post-season review examines score validity, question coverage, strategic behavior, concentration, participant burden, confidentiality, and downstream misuse before the next season is opened.","structural_mapping":[{"archetype_element":"Explicit rivalry purpose","domain_realization":"Use repeated probabilistic competition to identify forecasters who contribute independently informative estimates across assigned strategic questions, rather than to reward prediction theater or certainty."},{"archetype_element":"Scarce prize","domain_realization":"Four rotating seats per season on an eight-member funded strategic forecast council."},{"archetype_element":"Competitor eligibility boundary","domain_realization":"Trained forecasters qualify through conflict disclosure, confidentiality commitments, identity verification, scoring-rule orientation, and acceptance of balanced assignment and audit."},{"archetype_element":"Contest arena boundary","domain_realization":"Independent research and scheduled probability updates are allowed; copying sealed forecasts, proxy participation, privileged-data misuse, resolver lobbying, harassment, resource-cap evasion, and attempts to influence outcomes are prohibited."},{"archetype_element":"Performance metric and scoring basis","domain_realization":"A precommitted proper score, balanced question assignment, stratum normalization, minimum completion requirements, and a rolling multi-season record connect rank to performance across comparable exposures."},{"archetype_element":"Fair process and due process","domain_realization":"Question writers and resolvers are separated from contestants, assignment is randomized within balanced blocks, rules are frozen, audits are documented, and independent appeals cover assignment, resolution, scoring, and eligibility errors."},{"archetype_element":"Anti-sabotage and anti-collusion guardrail","domain_realization":"Sealed forecasts, communication rules, identity checks, similarity and timing screens, investigation referrals, and graduated penalties address copying, coordinated ranking, proxy accounts, resolver interference, and outcome manipulation."},{"archetype_element":"Externality and spillover boundary","domain_realization":"Conflict recusals, nonpublication defaults, exclusion of readily manipulable or harmful questions, aggregate-only decision feeds, and explicit prohibition on automatic action limit effects on forecast subjects and nonparticipants."},{"archetype_element":"Escalation and arms-race damper","domain_realization":"Equal research-tool credits, common data access, fixed update counts, and bounded question loads limit expenditure and constant-update races."},{"archetype_element":"Winner power and lock-in review","domain_realization":"Staggered terms, seasonal promotion and relegation, no rulemaking role for council members, delayed standings, and recurring qualification prevent one victory from becoming permanent control."},{"archetype_element":"Learning and recalibration","domain_realization":"Post-season analysis tests question balance, score behavior, information contribution, burden, gaming, concentration, confidentiality, and decision use before revising or retiring the league."}],"mechanism_mapping":[{"mechanism_slug":"contest_rulebook","role":"Freezes eligibility, balanced assignment, sealed-play rules, scoring, recusals, resolution, audits, fouls, tie-breaks, promotion, and appeals before the season begins.","counterfactual_removal":"Contestants or administrators could reshape exposure, resolution, or promotion rules after observing who benefits, making rank neither comparable nor contestable."},{"mechanism_slug":"bracket_or_tournament_structure","role":"Organizes rivalry into balanced seasonal question blocks, a qualification record, an audited promotion round, staggered council terms, and recurring relegation rather than one undifferentiated lifetime leaderboard.","counterfactual_removal":"Easy-question exposure, one lucky run, or permanent accumulated rank could dominate selection without repeated comparison under a common competitive shape."},{"mechanism_slug":"ranked_leaderboard_with_audit","role":"Ranks contestants on precommitted proper scores across comparable assigned questions and audits the leaders' identity, timestamps, recusals, data access, resources, and independence before awarding seats.","counterfactual_removal":"A visible score could crown copied, selectively entered, misattributed, or resource-noncompliant forecasts and amplify gaming without verifying how the standing was produced."},{"mechanism_slug":"spending_cap_or_resource_cap","role":"Gives contestants equal research-tool credits, common data access, bounded question loads, and identical update allowances.","counterfactual_removal":"Rank could become a contest of analyst hours, purchased information, tool access, or continuous updating rather than forecasting judgment."},{"mechanism_slug":"multiple_award_or_portfolio_selection","role":"Promotes four forecasters at once while retaining four staggered members, with a concentration check across methods and conflicts so the council does not depend on one contestant or affiliated cluster.","counterfactual_removal":"A single champion or synchronized group could monopolize the advisory channel, making one forecasting failure or capture consequential for the whole council."},{"mechanism_slug":"sabotage_or_foul_penalty_schedule","role":"Precommits graduated consequences for sealed-entry sharing, proxy accounts, privileged-data misuse, resource evasion, resolver lobbying, harassment, and outcome manipulation.","counterfactual_removal":"Misconduct that improves rank could remain rational because prohibitions would lack known, consistently applicable consequences."},{"mechanism_slug":"anti_collusion_monitoring","role":"Applies uniform cross-round screens for improbable probability matching, synchronized updates, reciprocal information transfers, coordinated abstention, and stable seat rotation, referring flags to independent inquiry.","counterfactual_removal":"Contestants could manufacture independent-looking agreement or divide council seats while each account appears compliant when examined alone."},{"mechanism_slug":"challenger_access_window","role":"Reopens four council seats every season to qualified league entrants and requires incumbents whose terms expire to compete under the current rules.","counterfactual_removal":"Council access, internal information, and institutional familiarity could compound into permanent incumbent advantage despite the appearance of a forecasting competition."},{"mechanism_slug":"post_contest_impact_review","role":"Examines whether scoring, assignments, sealing, promotion, council aggregation, and decision use remained aligned with the league's purpose and feeds revisions into the next season.","counterfactual_removal":"The office could retain a leaderboard that had become predictable, collusive, burdensome, or disconnected from useful strategic estimates."}],"causal_chain":["Funded council seats create rivalrous dependence: one contestant's rank and promotion reduce the opportunities available to others.","Voluntary open forecasting lets contestants improve rank by selecting exposure, observing rivals, concentrating resources, or exploiting resolution rather than by contributing independent estimates across comparable uncertainties.","Balanced assignment and mandatory completion remove easy-question selection and strategic abstention as ordinary winning routes.","Sealed forecasts and fixed update windows prevent contestants from copying visible consensus or timing updates solely around rivals' disclosures.","Equal research credits and bounded workloads dampen expenditure and constant-update arms races.","Proper scoring across comparable strata, a rolling multi-season record, and leader audits couple promotion to verified performance rather than isolated wins or unaudited accounts.","Conflict recusals, foul enforcement, and anti-collusion referrals constrain outcome influence, proxy play, resolver capture, and coordinated seat allocation.","Multiple staggered seats preserve continuity and methodological plurality while recurring promotion windows stop incumbency from becoming ownership of the arena.","Post-season evidence triggers rule revision or league retirement if strategic adaptation disconnects rank from the intended forecasting contribution."],"baseline":"The office invites trained analysts to forecast whichever open questions interest them and publishes a live leaderboard. Participants face unequal question counts, difficulty, research time, tools, and update frequency. They can see one another's estimates, question authors may also compete, recusals are informal, resolution disputes are handled case by case, leaders are not routinely audited, and high-ranked forecasters are appointed to the council without fixed terms or a recurring challenger path.","nearest_rivals":["Expert appointment by managers avoids a tournament but makes council access depend on reputation and managerial judgment rather than comparable repeated forecasting behavior.","An open forecasting leaderboard measures resolved predictions but permits self-selected exposure, visible herding, unequal resources, and permanent accumulated advantage unless its rivalry is bounded.","A prediction market uses prices and trading incentives to aggregate beliefs but allows capital, liquidity, strategic trading, and market impact to shape influence rather than assigning comparable forecast blocks to individual contestants.","A Delphi panel iteratively moves experts toward a group judgment; it suppresses some interpersonal dominance but does not use independently scored rivalry to allocate scarce rotating council seats.","A simple forecast average diversifies estimates without governing who enters, how forecasts are produced, whether contributions are independent, or who receives continuing advisory authority.","A one-time forecasting exam supplies common questions but does not test repeated updating, long-horizon calibration, conflict handling, collusion, or incumbent contestability."],"remaining_contrastive_claim":"The proposal governs a longitudinal outcome-resolved rivalry among forecasters through balanced assignment, sealed play, equal research resources, proper scoring, promotion and relegation, and staggered council access. Its operative path is not comparative judging of foresight submissions, allocation of verification attention, or repair of provider infrastructure; it changes how repeated forecast performance produces and relinquishes scarce advisory authority.","authority_safety":{"decision_authority":"The strategic-foresight director may administer a no-stakes pilot and, with ethics and personnel approval, later authorize the league. The independent resolution committee controls question outcomes under the frozen rulebook. Council forecasts remain advisory; operational decision owners retain all authority over policy, spending, personnel, and external communications.","authorized_first_step":"Run a no-stakes microseason with volunteer or synthetic contestant identities, an equal research allowance, sealed forecasts, balanced assignment of a small set of approved non-sensitive questions, and no council appointments. Test question allocation, sealing, scoring, resolution, audit, and appeal workflows without using results for employment or operational decisions.","excluded_actions":["Using pilot standings for hiring, promotion, compensation, discipline, contracting, or public reputation","Allowing contestants to forecast questions whose outcomes they can materially influence","Publishing individual forecasts, identities, or sensitive aggregates without separate authorization","Automatically executing a decision because the council aggregate crosses a threshold","Treating statistical similarity or synchronized timing as proof of collusion","Changing question definitions, scoring, or promotion thresholds after forecasts are visible","Granting extra research credits, questions, updates, or resolver access to favored contestants","Penalizing a disclosed conflict recusal as an incorrect forecast","Permitting council members to write rules, resolve questions, or audit their own league records"],"halt_rollback":"Halt the pilot if forecasts or identities leak before sealing ends, a contestant can influence an assigned outcome, question definitions cannot be resolved under the frozen text, resource equality cannot be enforced, protected information is exposed, or an administrator changes scoring after viewing results. Void standings, make no appointments or personnel use of the data, revoke pilot access, quarantine sensitive records, notify participants, and return to the existing advisory process pending independent review."},"negative_tests":{"strongest_counterevidence":"The strongest counterevidence would show that the existing council-selection process already assigns balanced compulsory question sets, seals forecasts, equalizes research resources and updates, separates contestants from question writing and resolution, applies proper scoring across comparable strata, audits leaders, enforces conflict and anti-collusion rules, rotates seats, and reviews downstream use. That would leave little structural difference for the proposal to introduce.","problem_falsifier":"The problem is falsified if council seats are not scarce, contestants do not strategically select questions or observe rivals, research resources are already equivalent, participants cannot affect one another's rank or resolution, and verified records show that copying, abstention, resource disparity, collusion, outcome influence, and incumbent accumulation do not affect selection.","intervention_falsifier":"The intervention is undermined if balanced question blocks cannot be constructed, outcomes remain too ambiguous or delayed to score, normalization reverses rankings arbitrarily, sealed play prevents legitimate information synthesis without yielding interpretable independence, research caps cannot be audited, multi-season promotion excludes capable challengers, or performance in the league does not produce estimates usable by decision teams.","risks":["Proper scores may become targets and encourage probability adjustments that improve rank without improving decision usefulness.","Question writers may unintentionally favor familiar domains, horizons, or forecasting styles.","Normalization across question strata may introduce discretionary choices that determine promotion.","Sealed play may sacrifice useful collective learning and cause redundant research.","Research caps may be evaded through prior knowledge, informal assistance, or unrecorded tools.","Statistical collusion screens may flag legitimate convergence on common evidence.","Rolling multi-season records may stabilize noise but also delay access for strong new challengers.","Shorter-horizon resolved questions may dominate because long-range outcomes mature slowly.","Contestants may avoid candid extreme estimates if individual reputational consequences persist despite delayed standings.","Outcome resolution may embed political or interpretive judgments.","Council members may acquire privileged contextual information that advantages them when they reenter the league.","The council aggregate may be treated as authoritative beyond the questions and uncertainty it represents.","Competitive ranking may reduce voluntary information sharing or create participant stress.","Multiple seats may still be captured by an affiliated cluster using superficially distinct accounts or methods." ]},"next_evidence_step":"Conduct one preregistered no-stakes microseason using a bounded set of non-sensitive questions with observable intermediate signposts. Record assignment balance, completion, recusals, seal integrity, resource-credit compliance, update timing, score sensitivity to normalization, evaluator agreement on resolution, audit corrections, forecast-correlation flags, appeals, participant burden, and the interpretability of the resulting aggregate. Use the evidence only to assess mechanism feasibility and failure modes; make no council appointment or effect claim from the pilot.","prior_art_status":"UNSEARCHED","diversity_from_prior_proposals":"Proposal 1 governs judged scenario submissions competing for scarce preparedness-rehearsal slots through masked packages, hidden tests, and portfolio selection. This replacement does not select scenarios or rehearsals: it governs repeated outcome-resolved forecasts through randomized balanced exposure, sealed estimates, proper scoring, and promotion and relegation. Proposal 2 repairs incumbent vendor lock-in by moving data and interfaces into a protected common layer and operating production and challenger service lanes. This replacement neither procures a forecasting platform nor changes infrastructure ownership; it governs individual competitive behavior and the conversion of verified seasonal performance into temporary advisory seats. Proposal 3 rations upstream weak-signal nominations with nontransferable escalation warrants and confidential priority registration. This replacement does not allocate verification attention or reward signal discovery; every contestant receives an assigned question block, and rivalry turns on independently scored forecasts after resolution. Its problem, intervention, scarce prize, and causal path are independently adoptable from all three current proposals.","revision_record":{"parent_version":null,"progress_targets_addressed":["Replace non-distinct proposal 4","Preserve proposal_index 4 and version 0","Create an independently adoptable fourth opportunity","Avoid the judged foresight-submission intervention used by proposal 1","Provide complete authority, safeguards, rivals, falsifiers, and bounded evidence","Explain diversity from proposals 1, 2, and 3"],"conceptual_changes":["Replaced the planning-premise tribunal with a longitudinal forecasting promotion-and-relegation league","Changed the contested object from an official scenario-like premise to temporary membership on a standing forecast council","Changed the causal path from committee judging of submissions to balanced question exposure, sealed forecasts, proper outcome scoring, and recurring seat turnover"],"operational_changes":["Introduced randomized balanced question blocks, mandatory completion, sealed play, equal research credits, scheduled updates, multi-season scoring, staggered seats, leader audits, and outcome-resolution appeals","Removed identity-masked premise submissions, hidden model stress tests, paired advocacy rebuttal, and central-case portfolio designation"],"evidence_changes":["Prior art remains unsearched","Replaced the retrospective premise tribunal with a no-stakes prospective microseason focused on mechanism feasibility"],"claim_changes":["Makes no claim of novelty, prevalence, demand, or effect size","Narrows the contrastive claim to governance of repeated outcome-resolved forecasting rivalry"]}}