feat(event): score CHRONOS occurrence forecasts with a Brier rule

ARCHITECTURE.md

@@ -53,7 +53,7 @@ boundaries above remain the target modular MSA architecture.
 |---|---|
 | `evidence_core` | immutable evidence domain primitives |
 | `temporal_core` | typed clocks, intervals, and temporal reasoning |
-| `event_core` | event instances, mentions, roles, and provenance |
+| `event_core` | event instances, mentions, roles, provenance, and CHRONOS occurrence-prediction calibration |
 | `relation_graph` | typed relations and forward-transition validation |
 | `membership_core` | time-varying cross-classified multiple membership |
 | `persistence_postgres` | PostgreSQL repositories and migrations |

CHANGELOG.md

@@ -6,6 +6,7 @@ All notable changes to TEPP are documented here. The format follows Keep a Chang
 
 ### Added
 
+- `event_core` CHRONOS occurrence-prediction calibration: forecasts stay hypothetical, refuse promotion to event instances, and recover a computed Brier score against later-observed occurrence truth, with empty or mismatched streams failing closed.
 - `tepp_api` adaptive orchestration router (ADR 0010): versioned `direct`/`verify`/`committee`/`conductor`/`abstain` selection from CPU `f64` risk, ambiguity, evidence, and token-budget inputs; recorded stages, recursion, decomposition, access lists, and role-specific reasoning effort; fail-closed document-controlled policy/access/credentials; LLM plans remain proposals under deterministic statistical authority; comparable-budget ablation requires a direct baseline; credential-free contextual-orchestrator binding. Live NIM HTTP remains accepted-target.
 - `tepp_api` purpose-bound provider-payload minimization: time-bounded `PurposeGrant` evaluation, fail-closed expired/not-yet-valid/inverted/cross-tenant/impossible-calendar denial, semantic UTC calendar validation, refusal to copy identity mappings into model-provider payloads or ordinary logs, preservation of opaque analytical identifiers and membership roles (no blanket PII mask), a separately authorized scientific re-identification path, and an internally bound FIPS 180-4 SHA-256 audit digest appended through `ReidentificationAuditSink` before disclosure.
 - `persistence_postgres` backup/restore integrity: restored snapshots stay unusable until tenant, canonical `SHA-256`, knowledge-cutoff eligibility, temporal window order, and append-only triggers revalidate; SQL probes raise `restore integrity failed` (ADR 0013).

DOCUMENTATION.md

@@ -33,6 +33,7 @@ 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) |
+| CHRONOS prediction-calibration doctoring | [`docs/research/chronos-prediction-calibration.md`](docs/research/chronos-prediction-calibration.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) |
 | Adaptive orchestration router doctoring | [`docs/research/adaptive-orchestration-router.md`](docs/research/adaptive-orchestration-router.md) |

crates/event_core/src/error.rs

@@ -20,6 +20,10 @@ pub enum EventError {
     UnsupportedWireVersion,
     /// An unknown event-role name was supplied.
     UnknownEventRole,
+    /// A CHRONOS occurrence prediction was treated as an event instance.
+    PredictionIsNotEventInstance,
+    /// An unknown occurrence-truth label was supplied.
+    UnknownOccurrenceTruth,
 }
 
 impl fmt::Display for EventError {
@@ -32,6 +36,8 @@ impl fmt::Display for EventError {
             Self::InvalidWirePayload => "invalid event wire payload",
             Self::UnsupportedWireVersion => "unsupported event wire version",
             Self::UnknownEventRole => "unknown event role",
+            Self::PredictionIsNotEventInstance => "CHRONOS prediction is not an event instance",
+            Self::UnknownOccurrenceTruth => "unknown occurrence truth label",
         };
         formatter.write_str(message)
     }
@@ -65,6 +71,14 @@ mod tests {
                 "unsupported event wire version",
             ),
             (EventError::UnknownEventRole, "unknown event role"),
+            (
+                EventError::PredictionIsNotEventInstance,
+                "CHRONOS prediction is not an event instance",
+            ),
+            (
+                EventError::UnknownOccurrenceTruth,
+                "unknown occurrence truth label",
+            ),
         ] {
             assert_eq!(error.to_string(), message);
         }

crates/event_core/src/lib.rs

@@ -4,13 +4,15 @@
 //!
 //! TEPP separates **fallible event mentions** grounded in evidence from
 //! **versioned event instances** used for temporal state, multilevel membership,
-//! and scientific estimation. Mentions never silently become instances.
+//! and scientific estimation. Mentions and CHRONOS occurrence forecasts never
+//! silently become instances.
 
 mod confidence;
 mod error;
 mod identifier;
 mod instance;
 mod mention;
+mod prediction;
 mod registry;
 mod role;
 
@@ -30,6 +32,16 @@ pub use instance::EventInstance;
 pub use instance::refuse_mention_as_instance;
 /// Fallible textual event mention.
 pub use mention::EventMention;
+/// One CHRONOS occurrence forecast that remains hypothetical.
+pub use prediction::ChronosOccurrenceForecast;
+/// Opaque CHRONOS occurrence-prediction identity.
+pub use prediction::ChronosPredictionId;
+/// Later-observed occurrence truth for a CHRONOS forecast.
+pub use prediction::OccurrenceTruth;
+/// Mean squared error of CHRONOS occurrence forecasts against later truth.
+pub use prediction::chronos_prediction_brier_score;
+/// Explicit refusal to treat a CHRONOS prediction as an event instance.
+pub use prediction::refuse_prediction_as_instance;
 /// In-memory registry separating mentions from instances.
 pub use registry::EventRegistry;
 /// Typed event role kind.

crates/event_core/src/prediction.rs

@@ -0,0 +1,207 @@
+//! CHRONOS occurrence forecasts stay hypothetical until later evidence.
+
+use crate::{EventConfidence, EventError, EventInstanceId};
+
+/// Opaque CHRONOS occurrence-prediction identity.
+///
+/// A forecast is hypothesized future or schema-completion evidence. It is
+/// never a promoted event instance.
+#[derive(Clone, Copy, Debug, Eq, Hash, Ord, PartialEq, PartialOrd)]
+pub struct ChronosPredictionId(u32);
+
+impl ChronosPredictionId {
+    /// Reconstruct a prediction identity from a raw fixture or estimator label.
+    #[must_use]
+    pub const fn from_raw(raw: u32) -> Self {
+        Self(raw)
+    }
+
+    /// Return the raw prediction label.
+    #[must_use]
+    pub const fn raw(self) -> u32 {
+        self.0
+    }
+}
+
+/// Later-observed occurrence truth for a CHRONOS forecast.
+///
+/// Truth is recovered from later evidence. It does not rewrite the forecast
+/// into an event instance.
+#[derive(Clone, Copy, Debug, Eq, PartialEq)]
+pub enum OccurrenceTruth {
+    /// Later evidence established that the predicted event occurred.
+    Occurred,
+    /// Later evidence established that the predicted event did not occur.
+    DidNotOccur,
+}
+
+impl OccurrenceTruth {
+    /// Return the stable wire label name.
+    #[must_use]
+    pub const fn wire_name(self) -> &'static str {
+        match self {
+            Self::Occurred => "occurred",
+            Self::DidNotOccur => "did_not_occur",
+        }
+    }
+
+    /// Parse a stable wire occurrence-truth label.
+    ///
+    /// # Errors
+    ///
+    /// Returns [`EventError::UnknownOccurrenceTruth`] for unrecognized names.
+    pub fn from_wire_name(name: &str) -> Result<Self, EventError> {
+        match name {
+            "occurred" => Ok(Self::Occurred),
+            "did_not_occur" => Ok(Self::DidNotOccur),
+            _ => Err(EventError::UnknownOccurrenceTruth),
+        }
+    }
+
+    /// Return whether later evidence established occurrence.
+    #[must_use]
+    pub const fn occurred(self) -> bool {
+        matches!(self, Self::Occurred)
+    }
+
+    /// Return the binary probability target used for Brier scoring.
+    ///
+    /// Occurred truth is `1.0`; non-occurrence is `0.0`.
+    #[must_use]
+    pub const fn as_probability_target(self) -> f64 {
+        match self {
+            Self::Occurred => 1.0,
+            Self::DidNotOccur => 0.0,
+        }
+    }
+}
+
+/// One CHRONOS occurrence forecast that remains hypothetical.
+#[derive(Clone, Copy, Debug, PartialEq)]
+pub struct ChronosOccurrenceForecast {
+    prediction_id: ChronosPredictionId,
+    probability: EventConfidence,
+}
+
+impl ChronosOccurrenceForecast {
+    /// Bind a prediction identity to an occurrence probability.
+    #[must_use]
+    pub const fn new(prediction_id: ChronosPredictionId, probability: EventConfidence) -> Self {
+        Self {
+            prediction_id,
+            probability,
+        }
+    }
+
+    /// Return the prediction identity.
+    #[must_use]
+    pub const fn prediction_id(self) -> ChronosPredictionId {
+        self.prediction_id
+    }
+
+    /// Return the hypothesized occurrence probability.
+    #[must_use]
+    pub const fn probability(self) -> EventConfidence {
+        self.probability
+    }
+}
+
+/// Explicit refusal to treat a CHRONOS occurrence prediction as an event instance.
+///
+/// # Errors
+///
+/// Always returns [`EventError::PredictionIsNotEventInstance`].
+pub fn refuse_prediction_as_instance(
+    _prediction: ChronosPredictionId,
+) -> Result<EventInstanceId, EventError> {
+    Err(EventError::PredictionIsNotEventInstance)
+}
+
+/// Mean squared error of CHRONOS occurrence probabilities against later truth.
+///
+/// # Errors
+///
+/// Returns [`EventError::InvalidWirePayload`] when the slices are empty or
+/// have unequal length.
+pub fn chronos_prediction_brier_score(
+    forecasts: &[ChronosOccurrenceForecast],
+    outcomes: &[OccurrenceTruth],
+) -> 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 residual = forecast.probability().value() - outcome.as_probability_target();
+        square_sum += residual * residual;
+    }
+    mean_square(square_sum, forecasts.len())
+}
+
+fn mean_square(square_sum: f64, count: usize) -> Result<f64, EventError> {
+    let n = u32::try_from(count).map_err(|_| EventError::InvalidWirePayload)?;
+    if n == 0 {
+        return Err(EventError::InvalidWirePayload);
+    }
+    Ok(square_sum / f64::from(n))
+}
+
+#[cfg(test)]
+mod tests {
+    use super::{
+        ChronosOccurrenceForecast, ChronosPredictionId, OccurrenceTruth,
+        chronos_prediction_brier_score, refuse_prediction_as_instance,
+    };
+    use crate::{EventConfidence, EventError};
+
+    #[test]
+    fn prediction_helpers_cover_local_branches() {
+        let prediction = ChronosPredictionId::from_raw(9);
+        assert_eq!(prediction.raw(), 9);
+        assert_eq!(
+            refuse_prediction_as_instance(prediction),
+            Err(EventError::PredictionIsNotEventInstance)
+        );
+        assert_eq!(OccurrenceTruth::Occurred.wire_name(), "occurred");
+        assert_eq!(OccurrenceTruth::DidNotOccur.wire_name(), "did_not_occur");
+        assert_eq!(
+            OccurrenceTruth::from_wire_name("occurred").expect("parse"),
+            OccurrenceTruth::Occurred
+        );
+        assert_eq!(
+            OccurrenceTruth::from_wire_name("did_not_occur").expect("parse"),
+            OccurrenceTruth::DidNotOccur
+        );
+        assert_eq!(
+            OccurrenceTruth::from_wire_name("maybe"),
+            Err(EventError::UnknownOccurrenceTruth)
+        );
+        assert!(OccurrenceTruth::Occurred.occurred());
+        assert!(!OccurrenceTruth::DidNotOccur.occurred());
+        assert!((OccurrenceTruth::Occurred.as_probability_target() - 1.0).abs() < f64::EPSILON);
+        assert!((OccurrenceTruth::DidNotOccur.as_probability_target() - 0.0).abs() < f64::EPSILON);
+
+        let forecast = ChronosOccurrenceForecast::new(
+            prediction,
+            EventConfidence::new(0.25).expect("probability"),
+        );
+        assert_eq!(forecast.prediction_id(), prediction);
+        assert!((forecast.probability().value() - 0.25).abs() < f64::EPSILON);
+        let miss = chronos_prediction_brier_score(&[forecast], &[OccurrenceTruth::Occurred])
+            .expect("miss");
+        assert!((miss - 0.5625).abs() < 1e-15);
+        assert_eq!(
+            chronos_prediction_brier_score(&[], &[OccurrenceTruth::Occurred]),
+            Err(EventError::InvalidWirePayload)
+        );
+        assert_eq!(
+            super::mean_square(0.0, 0),
+            Err(EventError::InvalidWirePayload)
+        );
+        assert_eq!(
+            super::mean_square(1.0, usize::MAX),
+            Err(EventError::InvalidWirePayload)
+        );
+        assert!((super::mean_square(1.0, 2).expect("half") - 0.5).abs() < f64::EPSILON);
+    }
+}

crates/event_core/tests/prediction_calibration_contract.rs

@@ -0,0 +1,59 @@
+//! CHRONOS occurrence forecasts stay hypothetical and recover Brier scores.
+
+use event_core::{
+    ChronosOccurrenceForecast, ChronosPredictionId, EventConfidence, EventError, OccurrenceTruth,
+    chronos_prediction_brier_score, refuse_prediction_as_instance,
+};
+
+fn forecast(raw: u32, probability: f64) -> ChronosOccurrenceForecast {
+    ChronosOccurrenceForecast::new(
+        ChronosPredictionId::from_raw(raw),
+        EventConfidence::new(probability).expect("probability"),
+    )
+}
+
+#[test]
+fn chronos_prediction_cannot_be_cast_to_an_instance() {
+    let prediction = ChronosPredictionId::from_raw(7);
+    assert_eq!(
+        refuse_prediction_as_instance(prediction),
+        Err(EventError::PredictionIsNotEventInstance)
+    );
+}
+
+#[test]
+fn perfect_occurrence_forecasts_recover_zero_brier() {
+    let forecasts = [forecast(1, 1.0), forecast(2, 0.0), forecast(3, 1.0)];
+    let outcomes = [
+        OccurrenceTruth::Occurred,
+        OccurrenceTruth::DidNotOccur,
+        OccurrenceTruth::Occurred,
+    ];
+    let score = chronos_prediction_brier_score(&forecasts, &outcomes).expect("brier");
+    assert!(score.abs() < 1e-15, "perfect Brier {score}");
+}
+
+#[test]
+fn calibrated_forecasts_beat_overconfident_always_occur_and_mismatches_fail_closed() {
+    let calibrated = [forecast(1, 0.8), forecast(2, 0.2), forecast(3, 0.7)];
+    let overconfident = [forecast(1, 1.0), forecast(2, 1.0), forecast(3, 1.0)];
+    let outcomes = [
+        OccurrenceTruth::Occurred,
+        OccurrenceTruth::DidNotOccur,
+        OccurrenceTruth::Occurred,
+    ];
+    let calibrated_brier = chronos_prediction_brier_score(&calibrated, &outcomes).expect("cal");
+    let naive_brier = chronos_prediction_brier_score(&overconfident, &outcomes).expect("naive");
+    assert!(
+        calibrated_brier < naive_brier,
+        "calibrated Brier {calibrated_brier} must be below always-occur Brier {naive_brier}"
+    );
+    assert_eq!(
+        chronos_prediction_brier_score(&calibrated, &[OccurrenceTruth::Occurred]),
+        Err(EventError::InvalidWirePayload)
+    );
+    assert_eq!(
+        chronos_prediction_brier_score(&[], &[]),
+        Err(EventError::InvalidWirePayload)
+    );
+}

docs/LLM_ORCHESTRATION.md

@@ -1,6 +1,6 @@
 # TEPP LLM Orchestration and Test-Time Compute Contract
 
-**Status:** Partial — `tepp_api::route_orchestration` is the governed selector; live provider execution is not yet shipped.  
+**Status:** Partial — `tepp_api::route_orchestration` is the governed selector; live provider execution is not yet shipped.
 **Last reviewed:** 2026-08-13
 
 ## 1. Purpose
@@ -132,4 +132,4 @@ Before claiming an orchestration mode materially improves TEPP, compare at least
 4. adaptive/learned-conductor-style workflow where available;
 5. at least two reasoning-effort/budget settings.
 
-Report uncertainty and failure modes, not only the best benchmark score.
+Report uncertainty and failure modes, not only the best benchmark score.

docs/TRACEABILITY.md

@@ -1,7 +1,7 @@
 # TEPP Requirements, Research, and Evidence Traceability
 
 **Status:** Accepted cross-cutting traceability baseline  
-**Last reviewed:** 2026-08-13
+**Last reviewed:** 2026-08-20
 
 The full APA 7th standards/literature register remains `docs/research/standards-and-literature.md`. This matrix links durable requirements to their owning decisions and implementation/evidence maturity without duplicating the bibliography.
 
@@ -30,7 +30,7 @@ The full APA 7th standards/literature register remains `docs/research/standards-
 | compositional topic correlation / stable clustering | ADR 0005/0012; research | future `network_analysis` | accepted-target |
 | posterior ESEM / longitudinal invariance / DSEM | ADR 0005 | future `psychometric_core` | accepted-target |
 | CPU bounded multithreading + GPU/VRAM streaming/parity | ADR 0001/0006 | future `compute_backend` | accepted-target |
-| TDT detection/tracking vs CHRONOS schema/prediction/temporal consistency | ADR 0016; PRD/research | future `event_intelligence` | accepted-target |
+| TDT detection/tracking vs CHRONOS schema/prediction/temporal consistency | ADR 0016; PRD/research | `event_core` occurrence-prediction Brier calibration on the active PR; remaining detection/schema/temporal-consistency stack future | active-PR |
 | evidence-bounded LLM interpretation | ADR 0010/0012; PRD | `tepp_api` router plus future `interpretation_gateway` | partial |
 | adaptive direct/verify/committee/conductor test-time compute | ADR 0010; `docs/LLM_ORCHESTRATION.md` | `tepp_api::route_orchestration` + ablation record on the active PR; live contextual-orchestrator execution remaining | partial |
 | purpose-bound PII handling without blanket masking | ADR 0009; `docs/PRIVACY_DATA_GOVERNANCE.md` | `tepp_api` export authorization plus provider-payload minimization / elevated re-identification implemented-main; persistence retention/deletion remaining | partial |