feat(event): score mention confidence with a known-truth Brier rule

CHANGELOG.md

@@ -6,6 +6,7 @@ All notable changes to TEPP are documented here. The format follows Keep a Chang
 
 ### Added
 
+- `event_core` mention-confidence Brier score: known-truth binary outcomes recover a computed Brier of 0 for perfect forecasts and 0.25 for constant 0.5, with empty or mismatched streams failing closed.
 - `membership_core` nested ICC: CPU `f64` unbalanced ANOVA recovers a known cluster ICC and refuses to treat cross-classified or multiple-membership designs as a single hierarchy (ADR 0003).
 - `persistence_postgres` typed `text_segment` SQL: insert/lookup of exact UTF-8 half-open byte spans on the existing `0006` table, cutoff-eligible document reads (`available_time <= knowledge_cutoff`), and live recovery of a known `hello` span. No new migration number (`#45` still owns `0007`).
 - Hourly contextual-orchestrator discovery records all provider models but routes OpenCode only through general-chat candidates, excluding embedding, image, reranker, transcription, moderation, safety, and other endpoint-only identifiers before price selection.

DOCUMENTATION.md

@@ -34,6 +34,7 @@ TEPP's approved PRD v0.4 and implementation plan are the primary product baselin
 | Hourly NIM product-development operations | [`docs/operations/HOURLY_NIM_PRODUCT_DEVELOPMENT.md`](docs/operations/HOURLY_NIM_PRODUCT_DEVELOPMENT.md) |
 | Actions workflow fleet audit | [`docs/operations/ACTIONS_WORKFLOW_FLEET.md`](docs/operations/ACTIONS_WORKFLOW_FLEET.md) |
 | Actions fleet research doctoring | [`docs/research/actions-workflow-fleet.md`](docs/research/actions-workflow-fleet.md) |
+| Mention-confidence Brier doctoring | [`docs/research/mention-confidence-brier.md`](docs/research/mention-confidence-brier.md) |
 | Event-intelligence status-gate doctoring | [`docs/research/event-intelligence-status-gates.md`](docs/research/event-intelligence-status-gates.md) |
 | Retention/deletion/legal-hold doctoring | [`docs/research/retention-deletion-legal-hold.md`](docs/research/retention-deletion-legal-hold.md) |
 | Provider-payload minimization doctoring | [`docs/research/provider-payload-minimization.md`](docs/research/provider-payload-minimization.md) |

crates/event_core/src/confidence.rs

@@ -39,6 +39,29 @@ impl EventConfidence {
     }
 }
 
+/// Mean squared error of mention probabilities against binary truth.
+///
+/// # Errors
+///
+/// Returns [`EventError::InvalidWirePayload`] when the slices are empty or
+/// have unequal length.
+pub fn mention_brier_score(
+    forecasts: &[EventConfidence],
+    outcomes: &[bool],
+) -> Result<f64, EventError> {
+    if forecasts.is_empty() || forecasts.len() != outcomes.len() {
+        return Err(EventError::InvalidWirePayload);
+    }
+    let mut square_sum = 0.0_f64;
+    for (forecast, outcome) in forecasts.iter().zip(outcomes) {
+        let target = if *outcome { 1.0 } else { 0.0 };
+        let residual = forecast.value() - target;
+        square_sum += residual * residual;
+    }
+    #[allow(clippy::cast_precision_loss)]
+    Ok(square_sum / forecasts.len() as f64)
+}
+
 #[cfg(test)]
 mod tests {
     use super::EventConfidence;
@@ -56,5 +79,9 @@ mod tests {
             EventConfidence::new(f64::NAN),
             Err(EventError::InvalidEventConfidence)
         );
+        let one = EventConfidence::certain().expect("certain");
+        assert!((one.value() - 1.0).abs() < 1e-15);
+        let miss = super::mention_brier_score(&[one], &[false]).expect("miss");
+        assert!((miss - 1.0).abs() < 1e-15);
     }
 }

crates/event_core/src/lib.rs

@@ -19,6 +19,8 @@ mod role;
 
 /// Finite confidence on the closed unit interval.
 pub use confidence::EventConfidence;
+/// Mean squared error of mention probabilities against binary truth.
+pub use confidence::mention_brier_score;
 /// Fail-closed event-ontology errors.
 pub use error::EventError;
 /// Opaque event-instance identifier.

crates/event_core/tests/confidence_calibration_contract.rs

@@ -0,0 +1,37 @@
+//! Mention confidence recovers known Brier scores against binary truth.
+
+use event_core::{EventConfidence, EventError, mention_brier_score};
+
+#[test]
+fn perfectly_calibrated_forecasts_recover_zero_brier() {
+    let forecasts = [
+        EventConfidence::new(0.0).expect("0"),
+        EventConfidence::new(1.0).expect("1"),
+        EventConfidence::new(0.0).expect("0"),
+        EventConfidence::new(1.0).expect("1"),
+    ];
+    let outcomes = [false, true, false, true];
+    let score = mention_brier_score(&forecasts, &outcomes).expect("brier");
+    assert!(score.abs() < 1e-15, "perfect Brier {score}");
+}
+
+#[test]
+fn constant_half_recovers_quarter_and_mismatches_fail_closed() {
+    let forecasts = [
+        EventConfidence::new(0.5).expect("half"),
+        EventConfidence::new(0.5).expect("half"),
+    ];
+    let outcomes = [false, true];
+    let score = mention_brier_score(&forecasts, &outcomes).expect("half");
+    let residual = score - 0.25;
+    let rmse = (residual * residual).sqrt();
+    assert!(rmse < 1e-15, "Brier RMSE {rmse}");
+    assert_eq!(
+        mention_brier_score(&forecasts, &[true]),
+        Err(EventError::InvalidWirePayload)
+    );
+    assert_eq!(
+        mention_brier_score(&[], &[]),
+        Err(EventError::InvalidWirePayload)
+    );
+}

docs/TRACEABILITY.md

@@ -13,8 +13,8 @@ The full APA 7th standards/literature register remains `docs/research/standards-
 | six distinct clocks and uncertain intervals | PRD; ADR 0002; ISO 24617-1:2012; Hobbs & Pan (2017) | merged PR #8 `temporal_core` on protected main; PR #5 historical lineage only | implemented-main |
 | Allen relation algebra/bounded closure | ADR 0002; Allen (1983) | merged PR #9 `temporal_core` path-consistency on protected main | implemented-main |
 | forward-only transition subgraph | PRD; ADR 0002/0003 | `relation_graph` on protected main | implemented-main |
-| event ontology/evidence mentions | PRD; ADR 0003 | `event_core` mention/instance separation on protected main; `persistence_postgres` mention SQL implemented-main refuses mention-as-instance; event-instance SQL (#39 implemented-main) refuses inverted windows; full intelligence stack remaining | partial |
-| time-varying cross-classified multiple membership | PRD; ADR 0003 | `membership_core` network and Kish ESS on protected main; nested ICC + cross-classified/multiple-membership refusal (this PR); full multilevel/MMMC/ESEM estimators remaining | partial |
+| event ontology/evidence mentions | PRD; ADR 0003 | `event_core` mention/instance separation on protected main; `persistence_postgres` mention SQL implemented-main refuses mention-as-instance; event-instance SQL (#39 implemented-main) refuses inverted windows; Brier calibration on the active PR; full intelligence stack remaining | partial |
+| time-varying cross-classified multiple membership | PRD; ADR 0003 | `membership_core` network on protected main; multilevel estimators remaining | partial |
 | leakage-safe availability/cutoff snapshots | PRD; ADR 0002/0013 | `corpus_split` on protected main | implemented-main |
 | recovery metrics (RMSE, bias, coverage, graph, temporal order, Monte Carlo SE gates) | PRD; Test Strategy; ADR 0007/0014 | `validation_core` on protected main (PR #19); SE-aware Monte Carlo gates included | implemented-main |
 | PostgreSQL bitemporal/lineage persistence | ADR 0013; Architecture/ERD | `persistence_postgres` migration contracts, in-memory adapters, live SQL session/document SQL port, tenant RLS (`0002` + session GUC/role helpers), `DATABASE_URL` SQLx gate, optional `live-sqlx` `PgPool` driver, exact-head live PostgreSQL CI with isolation proof, append-only immutability triggers (`0004`), temporal interval ordering CHECKs (`0005`), typed membership assignment (`0006` implemented-main), event-relation/mention/instance SQL (#37–#39 implemented-main), source-artifact SQL (#40 implemented-main), audit-event SQL (#41 implemented-main), concurrent document-write stress (#43 implemented-main), backup/restore integrity revalidation (#44 implemented-main), typed `text_segment` SQL insert/cutoff lookup (active PR); remaining physical ERD constraints including `document_record` FK on `text_segment` | partial |

docs/research/mention-confidence-brier.md

@@ -0,0 +1,27 @@
+# Mention-confidence Brier score
+
+## Scope
+
+This note doctors the `event_core` calibration contract for fallible event mentions:
+
+1. mention confidence is a probability on `[0, 1]`;
+2. `mention_brier_score` is the mean squared error against binary truth;
+3. empty or length-mismatched streams fail closed.
+
+TDT/CHRONOS promotion remains on the event-intelligence active PR. No database migration is allocated.
+
+## Authoritative sources
+
+Brier, G. W. (1950). Verification of forecasts expressed in terms of probability. *Monthly Weather Review, 78*(1), 1–3. https://doi.org/10.1175/1520-0493(1950)078<0001:VOFEIT>2.0.CO;2
+
+Gneiting, T., & Raftery, A. E. (2007). Strictly proper scoring rules, prediction, and estimation. *Journal of the American Statistical Association, 102*(477), 359–378. https://doi.org/10.1198/016214506000001437
+
+## Application
+
+Brier (1950) defines the mean squared error of a probability forecast. Gneiting and Raftery (2007) treat the Brier score as a strictly proper scoring rule, so a mention that is certain when true and impossible when false is uniquely optimal. TEPP therefore scores mention confidence against known binary outcomes rather than treating a high score as an event instance (Brier, 1950; Gneiting & Raftery, 2007).
+
+## Verification
+
+- forecasts `(0,1,0,1)` against outcomes `(false,true,false,true)` recover Brier `0`;
+- constant `0.5` against mixed outcomes recovers `0.25` with computed residual RMSE;
+- empty and mismatched streams return `InvalidWirePayload`.

docs/validation/temporal-event-foundation.md

@@ -23,6 +23,7 @@ This report tracks exact-head scientific and engineering evidence required befor
 | Leakage-safe splits | `corpus_split` | implemented-main | — | cutoff + co-partition tests | Task 9 / PR #17 |
 | Truth corpora / manifests | `tepp_simulation` | implemented-main | — | deterministic generator tests | Task 10 / PR #18 |
 | Recovery metrics | `validation_core` | implemented-main | — | RMSE/bias/coverage/MC gates | Task 11 / PR #19 |
+| Mention-confidence Brier score | `event_core` | active-PR | calibration vs binary truth | perfect 0 / half 0.25 RMSE | ADR 0003; `docs/research/mention-confidence-brier.md` |
 | Checkpoint is not the estimator | `checkpoint_authority` | accepted-target | active PR | refuse checkpoint-as-estimator + unvalidated artifact + recovery vs estimator collapse | ADR 0001/0014 |
 | Versioned API/export contracts | `tepp_api` | implemented-main | naruon HTTP interchange | unknown-field/version/limit + naruon HTTPS interchange tests | Task 12 / PR #21; live HTTP service remaining |
 | TDT/CHRONOS evidence-status gates | `event_core` | active-PR | PR #50 | admission + first-story rates | known-stream miss/FA; full tracking/calibration/schema extraction remains future; ADR 0016; `docs/research/event-intelligence-status-gates.md` |