Struct Calibration 

pub struct Calibration {
    pub params: EstimateParams,
    pub missions_used: usize,
    pub doc_heavy_missions_used: usize,
    pub expected_mult: f64,
    pub low_mult: f64,
    pub high_mult: f64,
    pub excluded_non_frontier: usize,
    pub gate_correlation: Option<f64>,
}

EstimateParams derived from the repo’s completed missions, plus how many missions informed them. missions_used == 0 means the params are the built-in defaults (nothing to calibrate against yet).

Fields

params: EstimateParams

missions_used: usize

doc_heavy_missions_used: usize

How many of the completed missions folded into params classified as MissionShape::DocHeavy (from their recorded plan). Zero means the doc-heavy shape is uncovered by this corpus — see apply_shape.

expected_mult: f64

Multiplier on the raw estimate expected_usd that recenters it onto the corpus (aggregate actual ÷ predicted). The calibrated per-run params already carry most of the correction, so this removes only the residual aggregate bias left by the count formula (≈1.08 on this repo’s corpus). 1.0 below MIN_CALIBRATION_MISSIONS.

low_mult: f64

Multipliers on the raw expected_usd for the low/high bounds: the empirical p10 / p90 of actual ÷ predicted, but clamped so the band only ever WIDENS the built-in 0.5 / 2.5 guess (in-sample quantiles from a small corpus understate a fresh mission’s uncertainty). 0.5 / 2.5 below MIN_CALIBRATION_MISSIONS.

high_mult: f64

excluded_non_frontier: usize

Completed missions EXCLUDED from the corpus for not being frontier spend (MissionCostClass::Local or MissionCostClass::Mixed) — their $0-marginal or blended actuals would distort the frontier calibration. Visibility for the corpus-size readouts (ready.rs shows missions_used).

gate_correlation: Option<f64>

Pearson r between per-mission gate activity and the actual÷predicted ratio (gate activity = blocked + grant requests + fix features + resumes). This is the calibrate-block-resume-cycles measurement; None below MIN_CALIBRATION_MISSIONS or when either series is constant.

Trait Implementations

impl Clone for Calibration

fn clone(&self) -> Calibration

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for Calibration

impl Debug for Calibration

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for Calibration

fn eq(&self, other: &Calibration) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for Calibration