---
title: Move seasonlog against-adjusted consumers onto the opponent baseline and retire the rows
type: task
status: draft
description: >-
  Move the league app's matchup fields and filter-prop-pairings.mjs onto the game sample engine's
  curated opponent baseline, then retire the seasonlog *_AGAINST_ADJ rows and their unscheduled
  rolling variants, which can be recomputed from gamelogs.
base_uri: user:task/league/move-seasonlog-against-adjusted-consumers-onto-opponent-baseline.md
created_at: '2026-10-23T23:57:06.661Z'
entity_id: 7bc08ff0-aba3-4fa3-907d-87f2336560dd
owner_identity_uri: █████████████████████████
priority: Low
public_read: true
tags:
  - user:tag/league-xo-football.md
updated_at: '2026-10-23T23:57:06.661Z'
visibility_analyzed_at: '2026-11-05T23:49:44.947Z'
---

Move the league app's matchup fields and the prop-pairing filter onto the game sample engine's opponent baseline, then retire the seasonlog `*_AGAINST_ADJ` rows and their rolling variants.

## Why

- `scripts/generate-nfl-team-seasonlogs.mjs:478-530` writes `*_AGAINST_ADJ` rows: a hand-rolled, uncurated leave-one-out over a whole season. The rolling variants are never scheduled, and merge-only upserts leave orphan rows.
- The engine now computes the same over-average reading curated and leave-one-out: the opponent baseline stage, `query-game-sample.mjs --opponent_baseline` and `query-slate-opponent-baselines.mjs`, from [[█████████████-██████████████-█████-███-████-███████-██████████]].
- Two consumers keep the rows alive: the league app UI (`app/core/player-fields.js:81-164`) and `scripts/filter-prop-pairings.mjs:324-350`. That is app work, so the operator kept it out of that plan (2026-10-23).
- The rows hold nothing unique: every value can be recomputed from team gamelogs and plays, so retiring them loses no data.

## Notes

Draft. Plan with [[user:workflow/write/write-software-implementation-plan.md]]: decide which app fields read the engine's residual, and whether `filter-prop-pairings.mjs` reads a curated baseline or drops the adjustment.