Balance Targets & Acceptance Guards

These are the numeric rails that keep runs inside the intended “Oregon Trail” feel and the gates the tester enforces.

Global pacing targets

• Route length target: ~2,000–2,400 mi (boss gate at distance_required, default 2,100; overlays may raise victory miles).
• Day window: 84–180 days.
• Mean miles per day: 10–20.
• Travel ratio: ≥ 0.90 (guards block configs that drop below this).

Classic family expectations

• Boss reach: 30–50% of runs.
• Boss win: ~20–35%.
• Survival (any non-early wipe ending): 60–80%.
• Failure mix: no single family (vehicle, sanity, exposure, crossings) > 50% of failures over large samples.
• Aggressive: lower survival and more terminal crossings than Balanced.
• Conservative / ResourceManager: higher survival and boss reach but boss win should stay ≲ 40%.

Deep family expectations

• Same distance/duration/mpd bands as Classic.
• Higher per-run variance is allowed; means must still satisfy the Classic bands.
• Crossings/failures may be a bit harsher or stranger, but not arcade-short or ultra-long.

Guard rails (from journey config)

• guards.min_travel_ratio: Classic 0.90; overlays may adjust.
• guards.target_distance: Classic 2000.0; overlay can raise (e.g., 2400.0).
• guards.target_days_min/max: 84 / 180.
• Tester sweeps (dystrail-tester) validate these across many seeds; failing guards fails CI.

Crossing & breakdown reference numbers

• Classic crossings: pass 0.72, detour 0.16, terminal 0.12; detour days 1–3.
• Bribe: pass bonus 0.2, terminal penalty 0.2, diminishing returns 0.4 (Balanced overlay tweaks).
• Breakdown (Classic base): p = base 0.01 * pace_factor * weather_factor + beta 0.10 * wear_component.
• Overlay example (Balanced): base 0.015, beta 0.185; pace multipliers 0.94 / 1.08 / 1.28; weather multipliers 1.0 / 1.25 / 1.35 / 1.1 / 1.15.