THE SCORING ENGINE
Give raw data a result you can explain.
The scorer is the heart of StatLine. Adapters turn raw fields into canonical metrics, profiles define the priorities, and the engine produces structured output you can inspect.
Understand the input
Load an adapter that knows the schema. Map fields, compute derived metrics, and preserve the context the model needs.
Make priorities explicit
Choose weights and score profiles that reflect your problem. Shared batch context keeps comparisons grounded in the same cohort.
Inspect the outcome
Read profile scores, raw aggregates, buckets, components, and weights. Choose a lean output when your application only needs the result.
Inspect the result.
Then inspect the rules.
A score is useful when you understand its inputs and assumptions. Explore an actual engine result, then open the adapter and profiles behind it.
PUBLIC SCORING INTERFACE
(adapter: 'AdapterLike', data: 'Mapping[str, Any] | Iterable[Mapping[str, Any]]', *, mode: 'str' = 'auto', weights: 'WeightsArg' = None, weights_override: 'dict[str, float] | None' = None, penalties_override: 'dict[str, float] | None' = None, output: 'OutputArg' = None, profiles: 'Sequence[str] | None' = None, filters: 'dict[str, Any] | None' = None, context: 'dict[str, dict[str, float]] | None' = None, caps_override: 'dict[str, float] | None' = None) -> 'Row | Rows'THE SCORER, AT WORKv4.0.0rc6
Raw data
Adapter
Scorer
↗Result
01 / INPUT
PLAYER"GoatedRahim"PPG31.909APG14.909RPG2.273
eba.players18 required fields · 4 profiles
✓02 / EXPLAINABLE OUTPUT
GoatedRahim
Houston Hornets
99PRI SCORE
Source-verified example · no browser scoring