Add game result context and W/L markers to team charts

[Mattsface]

Goal

Show whether each plotted team game was a win or loss so the Hits, Strikeouts, Runs, and future comparison charts can put the statistical trend in game-result context.

Data model

Prefer persisting the opponent's run total rather than a redundant W/L string:

• add opponent_runs to the normalized/persisted team-game record
• derive the game result from runs vs opponent_runs
• handle an equal score explicitly if the source ever represents a completed tie; do not silently classify equality as a loss

Because existing rows predate this field, preserve legacy data honestly. A migration should not fabricate opponent scores. If the new column must be nullable for old rows, fresh normalization should still require a real opponent score whenever the upstream source provides one, and re-importing a team-season should backfill it.

Before implementation, confirm which existing MLB response/schedule structure is the authoritative source for the opponent score. Do not add a new network call if the data is already present in the responses used for ingestion.

Chart behavior

For each individual-game observation, expose the result without changing the meaning of the metric:

• distinguish wins and losses on raw game markers
• include Result: W / Result: L in hover text
• keep rolling averages, team season averages, and MLB averages unchanged
• do not encode W/L into the rolling or reference lines
• preserve readability and accessibility; do not rely on color alone if marker shape or another cue can distinguish outcomes

Apply the result treatment consistently to the team-game charts that exist on main at implementation time. Do not duplicate result logic separately inside each chart builder if one small typed/helper representation can carry the outcome cleanly.

Semantics

The result is descriptive context only. A win marker does not imply that the plotted metric caused the win.

Prefer deriving:

runs > opponent_runs  -> W
runs < opponent_runs  -> L
runs = opponent_runs  -> T / explicit tie state if encountered

Do not store both scores and a second persisted result value unless there is a concrete reason; that would create two sources of truth.

Preserve

• existing Hits, Strikeouts, Runs analytics
• rolling-window calculations
• MLB-average calculations and COMPLETE coverage semantics
• browser routes remain DB-only
• server-rendered FastAPI/Jinja architecture
• local Plotly rendering
• empty/missing-data behavior

Testing

Cover at minimum:

• normalization/persistence of opponent runs
• migration behavior for legacy rows
• re-import/backfill behavior
• W/L derivation from team and opponent scores
• equality handled explicitly
• chart markers/hover show the correct result
• legacy rows without opponent score remain honest rather than inventing a result
• rolling/team/MLB averages remain numerically unchanged
• zero MLB API calls during normal browser rendering

Out of scope

• win probability
• Pythagorean expectation
• run-expectancy models
• causal claims between a metric and game outcome
• standings/record calculations
• using W/L to filter or recompute the statistical trend

Completion

A reader can look at an individual game point on the team charts and immediately tell whether that game was a win or loss, backed by persisted score data rather than a guessed or duplicated result field.