andreagrandi / andreagrandi/draftomen
Update replay, audit, and backtest for multi-card picks
- Lenguaje dominante
- Python
- Estrellas
- 0
- Forks
- 0
- Merge medio
- 1 h 21 min
- PR fusionados (30 d)
- 188
Descripción
## Problem
Replay, audit, and backtest records store one chosen card and use Quick Draft 14/42 arithmetic. Pick-Two would lose one selected card, misstate progress, and make singular actual-versus-recommended comparisons ambiguous.
## Proposed change
Carry draft format and selected-card collections through replay output, audit records, and backtest reconstruction. Define recommendation-following semantics for one-card and two-card logical picks without adding pair-optimization.
## Acceptance Criteria
- [ ] Replay output identifies the draft format and includes every card selected at each coordinate.
- [ ] Audit records preserve multi-card choices without changing coordinate identity.
- [ ] Backtest reconstructs logical picks and pool growth without equating pick count to card count.
- [ ] Recommendation-following semantics are deterministic for one-card and two-card picks.
- [ ] Existing Quick Draft golden output is intentionally migrated and remains deterministic.
- [ ] Focused replay/audit/backtest tests and the changelog cover every format.
## Scope notes
No scoring-model changes, UI work, or historical-data pipeline changes.
Guía de contribución
No hay ninguna guía de contribución indexada para este repositorio
Línea de trabajo
Locate the replay, audit, and backtest record and reconstruction code, then inspect the existing Quick Draft golden-output and focused tests to trace how a single selected card is represented. Update the formats and deterministic recommendation-following behavior for one-card and two-card picks, preserving coordinate identity and pool growth; done means every acceptance criterion is covered by focused tests and the changelog.
Escrito por el modelo de indexación a partir del texto del issue.
Evaluación
- Stack tecnológico
- python
- Área
- backend, testing
- Tipo de issue
- Nueva funcionalidad
- Dificultad
- 4/5
- Tiempo estimado
- 3-5 días
- Estado de actividad
- Activo
- Claridad
- Bastante claro
- Aptitud para principiantes
- 45/100