[khiet]

Summary

ADR 0104 answers the four questions ADR 0103 left for after the build. It uses a 40-seed bot probe because the playtest was not held; you chose to record the probe now and play later.

 docs/adr/
+└── 0104-rain-keeps-its-starting-grip-and-dry-stays-the-default.md
 crates/sim/src/tuning.rs
-    // Until rain has been played (ADR 0103).
+    // Until rooms have played Rain (ADR 0104).
 tools/repo-checks/test/adr.test.ts
+  "rain-keeps-its-starting-grip-and-dry-stays-the-default",

Question (ADR 0103)	Decision	Evidence from the probe
Retune the rain factors?	Keep x0.92 / x0.92 / x0.80 / +0.05 rad	Karts on grass are 0 or 1 in every row
Ace bots' wet row?	Keep the dry table	Going from 30% to new tyres wins 10.8 s a lap wet against 9.6 s dry. The table holds 10.2, and its measuring test allows 1 s either way
Lap 3 of 3, the oval	Keep both as they are	Stops after the rain starts are unchanged; the oval is 72% wet at the flag
RACE_END_CAP_S	Keep it	Goldstone over 5 laps caps more karts in rain (19 against 8 when rain falls on lap 3)
Default pick	Dry, until rooms have played Rain	No playtest yet

Still open: whether worn tyres in the wet read as rain or as a broken kart on keyboard, and how lap 3 of 3 feels to a human.

Evidence

• Before: the repo-checks ADR test lists 103 decisions.
  After: vitest run in repo-checks passes, 252 tests, with ADR 0104 in DECISION_SLUGS.
• The probe was a throwaway sim test that is not in this PR. Its output is quoted in the ADR's Context section.

Merge Danger

Door: two-way

The PR changes docs and one comment. No factor, protocol or replay changes.

Blast Radius: none

Closes #346