variable_name stringlengths 8 34 | measure_family stringclasses 3
values | track stringclasses 8
values | mathematical_definition stringlengths 19 69 | indexing stringlengths 14 94 | observation_unit stringlengths 7 32 | measurement_unit stringlengths 3 34 | economic_definition stringlengths 51 119 | economic_mechanism stringlengths 27 91 | hypothesis_role stringlengths 51 117 | source_mapping stringlengths 43 131 | decision_time_availability stringclasses 4
values | evidence_class stringclasses 8
values | missing_value_rule stringlengths 39 143 | validation_rule stringlengths 46 134 | references stringlengths 129 440 | measurement_limitations stringlengths 53 138 | uma_constructability stringclasses 2
values | uma_source_tables stringlengths 19 92 | uma_non_null_observations int64 34 2.66M | notes stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
bond_raw | B | capital | B_i = effectiveBond_i | i=OOV2 proposal/dispute | request | raw token integer | Principal placed at risk by the proposal and, when challenged, the dispute. | Risk-bearing commitment and the private cost of an incorrect claim. | Higher B_i may deter frivolous proposals/challenges but raises entry and capital costs. | polygon_uma_request_rounds.effective_bond_raw; currency supplies the asset | T1 | event | Null when no proposal exists; a source value of 0 is verified zero. | Integer string; positive for proposed requests; reconcile with token transfers. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.1016/0022-0531(79)90031-0 | Request-level bond does not alone identify each actor's duration of exposure. | complete | polygon_uma_request_rounds; polygon_uma_token_flows | 41,826 | |
reward_configured_raw | B | incentive | R_i^cfg = request.reward_i | i=OOV2 request | request | raw token integer | Reward offered by the requester before a report is supplied. | Ex-ante incentive for costly information production/reporting. | A larger configured reward may attract participation but is not yet realized income. | polygon_uma_request_rounds.question_reward_raw (cross-check reward_raw) | T0 | contract_parameter | Null means the request parameter was not recovered; zero is a valid no-reward request. | question_reward_raw = reward_raw where both are populated. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.2307/3003320; https://doi.org/10.1287/moor.6.1.58 | Configured reward may later be refunded or rolled and must not be labelled paid. | complete | polygon_uma_request_rounds | 42,675 | |
reward_to_bond_ratio | B | incentive | rho_i = R_i^cfg / B_i for B_i > 0 | i=OOV2 proposed request | request | dimensionless ratio | Configured compensation relative to capital placed at risk. | Incentive intensity and risk-return trade-off. | Higher rho_i may affect proposal or dispute participation conditional on observable selection. | question_reward_raw / effective_bond_raw within the same asset and decimals | T1 | derived | Null when B_i is null or zero; never divide by zero or combine assets. | Both inputs share currency; ratio is nonnegative and recomputable from raw integers. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.2307/3003320; https://doi.org/10.1287/moor.6.1.58 | Not a causal incentive estimate; request composition and adapter versions confound comparisons. | complete | polygon_uma_request_rounds | 41,826 | |
dispute_decision | C | monitoring | D_i = 1[DisputePrice event observed before d_i] | i=OOV2 proposal | request | binary | Whether at least one actor paid the on-chain cost to challenge the proposal. | Costly monitoring and public-good provision. | Low dispute rates may reflect agreement, weak monitoring, or high challenge cost; they do not prove truth. | polygon_uma_request_rounds.dispute_tx; polygon_oov2_events.event='DisputePrice' | T1 | event | 0 only when the full challenge window is observed; otherwise right-censored/null. | One request has at most one canonical dispute event in OOV2; transaction and event agree. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.1016/0022-0531(79)90031-0; https://www.jstor.org/stable/1805... | Undisputed acceptance is a protocol outcome, not independent factual correctness. | complete | polygon_uma_request_rounds; polygon_oov2_events | 41,827 | |
proposal_upheld | C | outcome | U_i = 1[resolvedPrice_i = proposedPrice_i] for disputed i | i=disputed OOV2 request | dispute | binary | Whether the DVM-supported settlement agrees with the proposal. | Ex-post adjudication of the challenger/proposer conflict. | Conditional outcome for studying challenge precision, not population truth accuracy. | proposed_price_raw and resolved_price_raw; Grade-A DVM link where cross-chain | terminal_only | cross_chain_link | Null for undisputed, unresolved or non-exactly linked requests. | For Grade-A links, DVM, child-pushed and OOV2 settled prices must be consistent. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://eth.blockscout.com/api/v2/smart-contracts/0x004395edb43EFca9885CEdad51E... | Protocol resolution may differ from independent ground truth. | complete | polygon_uma_request_rounds; uma_polygon_ethereum_grade_a_links | 832 | |
mandatory_wait_seconds | B | timing | W_i = d_i - t_i^proposal | i=OOV2 proposal | proposal | seconds | Protocol-defined minimum period during which a proposal remains challengeable. | Institutional delay that enables monitoring but locks capital. | Separates designed waiting time from excess operational delay. | expiration_time minus ProposePrice.block_time | T1 | derived | Null when either timestamp is missing; never assume the default liveness. | W_i >= 0; compare to exact event payload and contract version. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md | The period can be request-specific and must not be filled from the current global default. | complete | polygon_uma_request_rounds; polygon_oov2_events | 41,827 | |
settlement_delay_seconds | C | timing | L_i = t_i^settle - t_i^request | i=OOV2 request | request | seconds | Elapsed calendar time from request initiation to settlement. | Total institutional and operational friction. | Longer delay increases uncertainty and may increase capital opportunity cost. | Settle.block_time - RequestPrice.block_time | terminal_only | derived | Right-censored at the fixed cutoff for unsettled requests. | Nonnegative; both timestamps come from canonical event logs. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md | Does not by itself identify which stage caused delay. | complete | polygon_oov2_events | 41,826 | |
excess_delay_seconds | C | timing | X_i = max(0, L_i - W_i) | i=settled OOV2 request with observed W_i | request | seconds | Delay beyond the protocol-defined challenge window. | Operational friction net of mandatory institutional waiting. | Separates design-imposed waiting from settlement execution latency. | settlement_delay_seconds - mandatory_wait_seconds | terminal_only | derived | Null unless both L_i and W_i are known. | X_i >= 0 and inputs trace to event timestamps. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md | For disputed requests, DVM adjudication is part of excess delay and should later be decomposed. | complete | polygon_oov2_events; polygon_uma_request_rounds | 41,826 | |
reward_paid_raw | C | outcome | R_i^paid = explicitReportReward_i + disputeWinnerReward_i | i=settled OOV2 request and economic winner | request | raw token integer | Non-principal compensation actually embedded in the verified settlement transfer. | Realized incentive payment. | Distinguishes promised reward from realized income. | explicit_report_reward_raw + dispute_winner_reward_raw; settlement transfer QC | terminal_only | token_flow | Null when unsettled or transfer evidence unavailable; zero is valid after exact decomposition. | principal_returned + reward_paid = gross_payout where protocol fees are handled by the rule; flow QC exact. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.2307/3003320; https://doi.org/10.1287/moor.6.1.58 | Request-level winner payment requires an actor-role table before individual net payoff analysis. | complete | polygon_uma_request_rounds; polygon_uma_request_flow_qc; polygon_uma_token_flows | 41,826 | |
principal_returned_raw | C | capital | P_i^ret = decomposed returned bond principal | i=settled OOV2 request and settlement recipient | request | raw token integer | Recovery of previously locked principal; not income. | Capital recovery and exposure closure. | Separating return of capital prevents payout from overstating rewards. | polygon_uma_request_rounds.principal_returned_raw | terminal_only | token_flow | Null when unsettled; zero only when exact decomposition proves no principal return. | Must be reconciled within the same transaction and asset; never counted as reward. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.1016/0022-0531(79)90031-0 | A request aggregate may combine multiple principal components; actor allocation must be explicit. | complete | polygon_uma_request_rounds; polygon_uma_token_flows | 41,826 | |
bond_forfeited_raw | C | penalty | F_i^bond = losing bond amount not returned | i=settled disputed OOV2 request | dispute | raw token integer | Principal lost by the economically losing side of a dispute. | Penalty enforcement and risk realization. | Measures realized downside rather than a configured slash parameter. | polygon_uma_request_rounds.bond_forfeited_raw | terminal_only | token_flow | Null outside a settled eligible dispute; verified zero only after complete decomposition. | Positive only for settled disputed requests; reconcile winner transfer and protocol fee. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.2307/3003320; https://doi.org/10.1287/moor.6.1.58 | Final fee and redistributed bond components must remain separately identified. | complete | polygon_uma_request_rounds; polygon_uma_token_flows | 834 | |
protocol_fee_raw | B | cost | C_i^fee = protocol fee transferred from disputed escrow | i=settled disputed OOV2 request | dispute | raw token integer | On-chain fee retained or routed by the mechanism during dispute settlement. | Mechanism operation cost paid from escrow. | Reduces private net payoff and may affect willingness to challenge. | polygon_uma_request_rounds.protocol_fee_raw and final_fee_forfeited_raw | terminal_only | token_flow | Null outside eligible settled disputes; do not infer from current fee parameters. | Reconcile with settlement transaction flows and economic rule id. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.1016/0022-0531(79)90031-0; https://www.jstor.org/stable/1805... | Protocol fee is not the off-chain social cost of adjudication. | complete | polygon_uma_request_rounds; polygon_uma_token_flows | 834 | |
gross_payout_raw | C | outcome | G_i = total settlement transfer to recipient | i=settled OOV2 request | request | raw token integer | Total transferred amount, including returned principal and reward components; it is not an economic reward. | Cash-flow reconciliation variable, not an economic reward measure. | Used only to verify decomposition; treating G_i as reward overstates incentives. | polygon_uma_request_rounds.gross_payout_raw; request flow QC | terminal_only | token_flow | Null when unsettled or transfer unavailable. | gross_payout_raw equals settlement_transfer_raw for all exact-QC settlements. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.2307/3003320; https://doi.org/10.1287/moor.6.1.58 | Cannot be compared as income without subtracting capital and fees. | complete | polygon_uma_request_rounds; polygon_uma_request_flow_qc | 41,826 | |
realized_payoff_raw | C | outcome | Pi_ia = rewards_ia - forfeitures_ia - fees_ia - gas_ia | i=episode, a=actor | actor-episode | raw asset units by asset | Actor-specific realized economic gain or loss excluding returned principal. | Private economic consequence of reporting, monitoring or challenging. | Outcome for incentive compatibility and economic-regret evaluation. | Requires role-resolved token flows plus gas receipts and reward/forfeiture decomposition | retrospective_only | derived | Null unless all included components share asset and actor attribution; never treat missing cost as zero. | Component sum reconciles to actor-address flows; principal return excluded. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.2307/3003320; https://doi.org/10.1287/moor.6.1.58 | Current request table is not fully actor-role normalized and gas coverage is incomplete. | partial | polygon_uma_request_rounds; polygon_uma_token_flows; receipt archives | 41,826 | |
capital_days_locked_raw | B | capital | K_ia = B_ia * (t_release - t_lock)/86400 | i=episode, a=actor | actor-episode | raw token units multiplied by days | Quantity of capital tied up over time in its native asset. | Liquidity constraint and capital opportunity cost. | Higher K may discourage participation even if expected payoff is positive. | bond lock/release token flows and their block timestamps | terminal_only | derived | Right-censored capital-days use cutoff duration and retain censor flag; never sum across assets. | Release time cannot precede lock time; asset address and decimals remain attached. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.1016/0022-0531(79)90031-0 | Request-level principal is available, but exact actor-specific lock timestamps require a normalized flow lifecycle. | partial | polygon_uma_token_flows; polygon_oov2_events | 41,826 | |
gas_cost_native_raw | B | cost | C_ia^gas = gasUsed_tx * effectiveGasPrice_tx | i=decision transaction, a=actor | actor-episode | native chain smallest unit | Direct blockchain execution cost paid by the acting account. | Participation and verification cost. | Higher execution costs can suppress monitoring and challenge activity. | transaction receipt gasUsed and effectiveGasPrice for proposal/dispute/settlement actions | T1 | transaction_receipt | Null when receipt is absent; never zero-fill missing receipts. | Recompute product from hexadecimal receipt fields and match transaction sender. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.1016/0022-0531(79)90031-0; https://www.jstor.org/stable/1805... | Current archives cover targeted dispute and settlement receipts, not every relevant action uniformly. | partial | raw Polygon receipt archives | 835 | |
verification_cost_usd | B | cost | C_ia^verify = C_ia^gas,USD + C_ia^investigation + C_ia^delay | i=monitoring decision, a=monitor | actor-episode | USD at documented timestamps | Total private cost of investigating and executing a verification action. | Costly information acquisition and public-good provision. | Challenge occurs only when expected benefit exceeds verification and risk cost. | gas receipts plus externally measured labor/API/time costs and timestamped FX conversion | retrospective_only | external_evidence | Null if any required cost component is unavailable; partial cost must be labelled partial. | Each USD component cites source and conversion time; no silent zero labor cost. | NIST AI RMF characteristics; economic cost-accounting design in project guidance; https://doi.org/10.1016/0022-0531(79)90031-0; https://www.jstor.org/stable/1805228 | Gas is measured and timestamp-converted for the strict cohort; investigation and labor remain scenario assumptions. | partial | trustworthy_ai_usd_economics; Polygon dispute receipts; historical Chainlink proxy state | 810 | |
monitoring_concentration_hhi | A | monitoring | HHI_t = sum_a s_at^2, s_at = challenges_at / sum_b challenges_bt | a=challenger, t=preregistered time window | protocol-time window | dimensionless [0,1] | Concentration of observed challenge activity among monitoring actors. | Public-good provision, specialization and monitoring fragility. | High concentration can indicate reliance on a small set of monitors, without identifying welfare effects. | disputer addresses from canonical DisputePrice events | retrospective_only | derived | Null when actor attribution or full window coverage is unavailable; no challenges is not HHI=0. | Shares sum to one within protocol, chain, asset and time window. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.1016/0022-0531(79)90031-0; https://www.jstor.org/stable/1805... | Address concentration is not entity concentration; sybils and delegated execution remain possible. | complete | polygon_uma_request_rounds; polygon_oov2_events | 835 | |
monitoring_top_k_share | A | monitoring | TopShare_t(k) = sum_{a=1}^k s_(a)t for preregistered k in {1,5,10} | a=challenger ordered by activity share; t=preregistered protocol-time window | protocol-time-window-k | dimensionless [0,1] | Share of observed challenge activity supplied by the k most active monitoring addresses. | Verification concentration and dependence on a small monitor set. | A high top-k share indicates operational reliance on few addresses, not necessarily few entities. | canonical DisputePrice disputer addresses aggregated within protocol, chain, asset and time window | retrospective_only | derived | Null when the complete window or actor attribution is unavailable; no challenges is undefined, not zero. | Shares use the same denominator as HHI; TopShare_t(k) is nondecreasing in k and at most one. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.1016/0022-0531(79)90031-0; https://www.jstor.org/stable/1805... | Addresses may be sybils, delegates or shared infrastructure; k must be reported with every value. | complete | polygon_uma_request_rounds; polygon_oov2_events | 835 | |
monitoring_effective_n | A | monitoring | N_eff,t = 1 / HHI_t | t=preregistered protocol-time window with at least one observed challenge | protocol-time window | effective address count | Inverse-HHI effective number of equally active monitoring addresses. | Verification breadth and fragility under unequal participation. | Lower effective N indicates that nominal address counts overstate effective monitoring breadth. | monitoring_concentration_hhi derived from canonical DisputePrice disputer shares | retrospective_only | derived | Null when HHI is null or the window contains no observed challenges; never divide by zero. | N_eff,t lies in [1,n_t] and recomputes exactly as the inverse of the released HHI. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.1016/0022-0531(79)90031-0; https://www.jstor.org/stable/1805... | Effective address count is not an independently resolved count of human or institutional entities. | complete | polygon_uma_request_rounds; polygon_oov2_events | 835 | |
monitoring_threshold_n | A | monitoring | N_t(tau) = min{k: sum_{a=1}^k s_(a)t >= tau}, tau in {0.5,0.8,0.9} | a=challenger ordered by activity share; t=window; tau=preregistered coverage threshold | protocol-time-window-threshold | address count | Minimum number of top monitoring addresses needed to supply a specified share of challenges. | Threshold dependence and operational concentration of costly verification. | Small N_t(tau) indicates that removing a few active addresses could sharply reduce observed monitoring. | canonical DisputePrice disputer shares within the same denominator used by HHI and top-k share | retrospective_only | derived | Null without a complete nonempty activity window; tau is part of the observation key and cannot be omitted. | N_t(tau) is an integer in [1,n_t] and is nondecreasing in tau. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.1016/0022-0531(79)90031-0; https://www.jstor.org/stable/1805... | This is an address-activity threshold, not a protocol security threshold or causal failure estimate. | complete | polygon_uma_request_rounds; polygon_oov2_events | 835 | |
expected_verifier_utility_usd | B | incentive | EU_i(a|I_i,s) = sum_o p_i(o|I_i) Payoff_i(a,o,s) - Cost_i(a,s) | i=decision episode; a in {Accept,Investigate,Challenge,Abstain}; s=preregistered cost scenario | decision-episode-action-scenario | USD | Expected private utility of a feasible verifier action under a declared information set and cost scenario. | Economic incentive compatibility under uncertain protocol outcomes and costly verification. | A challenge is privately attractive only when its expected token payoff exceeds Gas, capital and investigation costs. | decision-time calibrated probability; exact protocol payoff mapping; timestamped FX/Gas; registered investigation and APR scenarios | T1 | derived | Null unless probability, action-contingent payoff and every included cost component are available; scenario assumptions are never observations. | No outcome-only field enters I_i; probabilities sum to one; all action utilities use one registered scenario and USD timestamp policy. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://airc.nist.gov/airmf-resources/airmf/3-sec-characteristics/; https://doi... | Private scenario utility is not social welfare; investigation labor and capital APR remain assumptions. | partial | trustworthy_ai_usd_economics; trustworthy_ai_complete_task; historical Chainlink proxy state | 810 | |
dvm_signed_payoff_delta_raw | C | outcome | Delta_iar = signedSlashTokens_iar | i=DVM request, a=voter, r=round/request index | actor-episode | raw UMA integer, signed | Request-level accrued redistribution: positive for correct-vote redistribution and negative for wrong/no-vote slash. | Voting incentive, participation penalty and redistribution. | Connects voting behavior to request-specific economic accrual. | uma_dvm_voter_payoffs.signed_slash_delta_raw from VoterSlashed | terminal_only | state_delta | Null if request-voter event is missing; zero is a real emitted zero delta. | positive + zero + negative classifications are exhaustive; VoterSlashApplied is excluded from this sum. | https://eth.blockscout.com/api/v2/smart-contracts/0x004395edb43EFca9885CEdad51EC9fAf93Bd34ac; reports/contract_semantics_audit.md; https://doi.org/10.2307/3003320; https://doi.org/10.1287/moor.6.1.58 | It is accrued to unappliedSlash and is not the same as the later account-level stake mutation. | complete | uma_dvm_voter_payoffs | 2,663,684 | |
right_censored | C | outcome | C_i = 1[t_i^terminal is not observed by fixed cutoff T] | i=economic episode; T=fixed release cutoff | economic episode | binary | Whether the lifecycle remains without an observed terminal event at the fixed release cutoff. | Accountability realization and duration measurement under incomplete follow-up. | Retention of censored episodes prevents completed-only outcome and delay estimates. | terminal event timestamps, protocol-native open/closed state and fixed cutoff manifest | terminal_only | derived | Never infer zero from a missing terminal timestamp unless source coverage through T and finality checks pass. | C_i=0 requires a canonical terminal event at or before T; C_i=1 retains observed duration T-t_i^start. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.2307/3003320; https://doi.org/10.1287/moor.6.1.58 | Administrative censoring does not identify the eventual outcome and must not be treated as failure. | complete | polygon_uma_request_rounds; polygon_oov2_events; release cutoff manifest | 42,676 | |
dvm_applied_stake_delta_raw | C | outcome | A_ak = emitted unappliedSlash at VoterSlashApplied_k | a=voter, k=stake-update transaction | actor stake-update | raw UMA integer, signed | Net accumulated slash/redistribution actually applied to the voter stake balance. | Realized account-level balance mutation. | Supports stake reconciliation but not request-level causal attribution. | uma_dvm_staking_events VoterSlashApplied raw_data_words and emitted postStake | terminal_only | state_delta | Null when application has not occurred by cutoff; not zero. | pre-stake + applied delta = emitted postStake subject to zero floor; never add VoterSlashed again. | https://eth.blockscout.com/api/v2/smart-contracts/0x004395edb43EFca9885CEdad51EC9fAf93Bd34ac; reports/contract_semantics_audit.md; https://doi.org/10.2307/3003320; https://doi.org/10.1287/moor.6.1.58 | One application can aggregate multiple request-level accruals. | complete | uma_dvm_staking_events | 262,448 | |
independent_ground_truth_available | C | outcome | G_i = 1[independent, timestamped reference outcome passes validation] | i=oracle request | request | binary | Whether outcome correctness can be judged independently of the protocol's own settlement. | Separates protocol consensus from empirical truth. | Required before reporting accuracy or truthfulness rather than protocol agreement. | validated external source record linked to immutable request semantics and resolution rule | retrospective_only | external_evidence | 0 only after an explicit coverage audit; otherwise unavailable/null. | External source timestamp and extraction rule are preserved and independent of model output. | https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; https://doi.org/10.2307/3003320; https://doi.org/10.1287/moor.6.1.58 | Available for a deterministic Binance-candle subcohort; most requests still provide only protocol outcomes. | partial | trustworthy_ai_independent_truth; Binance finalized klines | 34 | |
token_only_action_regret_raw | C | trustworthy_ai | Regret_i^token = max(0, Pi_i^challenge) - Pi_i(a_agent) | i=resolved decision episode; a in {Accept,Challenge,Abstain} | decision episode | raw episode-token integer | Ex-post private token payoff forgone relative to the better of taking no token action and challenging. | Action quality under asymmetric bond/reward consequences. | Tests whether an action policy improves private token outcomes before Gas and investigation costs. | exact OOV2 winner reward, losing bond/final-fee forfeiture, protocol outcome, and registered agent action | retrospective_only | derived | Null for Investigate because its subsequent action and investigation cost are unobserved; never subtract MATIC Gas from USDC. | Computed within one token address and decimals; challenge payoff reconciles to exact settlement flows. | https://raw.githubusercontent.com/UMAprotocol/protocol/a16ee53125c433dfa4e29738b73d9069ff109c03/packages/core/contracts/optimistic-oracle-v2/implementation/OptimisticOracleV2.sol; https://github.com/UMAprotocol/uma-docs/blob/master/faqs.md; project next-stage guidance; https://airc.nist.gov/airmf-resources/airmf/3-sec-... | Partial private-payoff regret only: excludes Gas, labor, external harm, independent truth and USD welfare. | partial | trustworthy_ai_requirements_audit; polygon_uma_request_rounds; Polygon dispute receipts | 810 | |
economic_regret_usd | C | trustworthy_ai | Regret_i = max_a U_i(a | I_i) - U_i(a_agent | I_i) | i=agent decision episode, a in {Accept,Investigate,Challenge,Abstain} | decision episode | USD | Economic loss from the agent's chosen action relative to the best feasible action under the preregistered payoff model. | Decision quality under costly verification, uncertainty and abstention. | Evaluates whether predictive gains survive action costs and asymmetric errors. | decision-time evidence, realized protocol outcome, cost matrix, gas and capital costs | retrospective_only | derived | Null without a preregistered utility matrix and sufficient realized cost/outcome evidence. | No future fields in I_i; recompute all action utilities and preserve cost assumptions. | NIST AI Risk Management Framework; project next-stage guidance; https://airc.nist.gov/airmf-resources/airmf/3-sec-characteristics/ | Constructed as private-verifier scenario regret with measured FX/Gas/payoff; investigation cost and APR are assumptions, not observations. | partial | trustworthy_ai_usd_economics; historical Chainlink proxy state; registered cost scenarios | 810 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.