# Objective
Technical reference for the `Objective` trait's precision-weighted scoring and the
built-in `ConsensusObjective` combinator (`objective/` module).
## Precision-weighted scoring
Both the `Objective` trait and the `Selector` (see [selector.md](selector.md)) implement
precision-weighted scoring as specified in ADR-024. The `precision()` hook returns an
inverse-variance estimate for each candidate's score; the default is 1.0 (fully
trusted). The trait enforces no range: any finite value is used verbatim in
ranking, and a non-finite value falls back to 1.0. The effective ranking score is:
$$\text{effective} = \text{score} \times \text{precision}$$
Non-finite precision falls back to 1.0. This allows objectives derived from uncertain models
(e.g., embedding similarity) to discount their own scores without propagating NaN into
ranking.
## `ConsensusObjective`
Uses the geometric mean of sub-objective scores:
$$\text{score} = \exp\!\left(\frac{1}{n}\sum_{i=1}^{n}\ln s_i\right)$$
Any sub-score at or below zero causes the consensus to return `0.0` (not an error). Callers
relying on `ConsensusObjective` should ensure sub-objectives return strictly positive scores
for passing candidates.
## Errors
- `ObjectiveError::NoCandidates` — `select_deterministic` called with empty slice.
- `ObjectiveError::NoMatch` — no candidate passes the minimum score threshold.