fix(physics): preserve first-candidate scan semantics

Carry predicted position separately through exact record-ID chunks, apply magnetic rewrites only at records 6/153/154, and honor type-three/dynamic broadphase. Match raw 1000:c79c by resolving fixed candidate slot one rather than a later nearer record, and restore related capture-age, tilt, and stacked multiball scoring side effects.

Test Plan:
- bash original/tools/test_reconstructed_c.sh
- python3 original/tools/audit_reconstruction.py --require-complete
- cargo test --all-targets
- cargo clippy --all-targets --all-features -- -D warnings
- rumdl check original/C_RECONSTRUCTION_FINAL_AUDIT.md original/MECHANICS_PROGRESS.md tdkpin-rs/CHANGELOG.md tdkpin-rs/RECONSTRUCTION.md tdkpin-rs/README.md
- git diff --check

tdkpin-rs/CHANGELOG.md

@@ -10,6 +10,10 @@ and this project adheres to
 
 ### Fixed
 
+- Match central score/event side effects: every stored collision candidate
+  clears its ball slot's capture age, Tilt suppresses all score and rule-record
+  mutations, and active multiball doubles points before permanent Double can
+  double them again.
 - Scan static, capture, trigger, magnetic, and dynamic record ranges in exact
   ID order with a separately tracked prediction. Preserve the raw scanner's
   first-detected-candidate quirk even when a later record is geometrically