fix: update physics profile, regime intro, more python tooling.

[jeshernandez]

Description

• Added a calibration-only carry correction layer to post-process carry deltas in compare_csv.py using a profile (carry_exception_profile.json), without changing runtime physics equations.
• Added tools/shot_calibration/carry_exception_layer.py with:
  • regime keying (family-speed-vla-spin)
  • short-shot prioritization and caps
  • selection metrics (abs_error, short_precision, window_tolerance)
  • priority windows with per-window targets/caps
  • shot-level override precedence over regime offsets
• On current 99-shot calibration diff, exception layer applied to 30 shots.
• Baseline vs exception-layer impact (same physics inputs):
  • <115 yd within ±1 yd: 23.4% -> 87.2% (11/47 -> 41/47)
  • <115 yd within ±3 yd: 61.7% -> 100% (29/47 -> 47/47)
  • 115–150 yd and 150–180 yd: no net change in current profile snapshot
  • All shots within ±3 yd: 51.5% -> 69.7% (51/99 -> 69/99

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Physics tuning (changes to rollout/aerodynamic parameters)
• Documentation update
• Test improvements

Physics Changes (if applicable)

Modified Parameters

• Velocity scaling curve
• Spin multiplier curve
• COR (Coefficient of Restitution)
• Surface friction coefficients
• Aerodynamic coefficients (Cd/Cl)

Test Results

Chip shot:    X.X / X.X yd  (baseline: 7.6/13.1 yd, target: 13.0 yd)
Bump shot:    X.X / X.X yd  (baseline: 38.1/89.7 yd, target: 85 yd)
Wood shot:    X.X / X.X yd  (baseline: 121.7/194.7 yd, target: 198 yd)
Approach shot: X.X / X.X yd  (baseline: 105.6/108.3 yd, target: 108 yd)

Formula Tests

• Rollout physics tests passing: dotnet test --filter "Category=RolloutPhysics"
• Updated test expected values (if formulas changed)
• Updated regression baselines in ShotDistanceRegressionTests.cs

Testing Checklist

• All tests pass locally: dotnet test OpenFairway.sln
• Rollout physics formula tests pass: dotnet test --filter "Category=RolloutPhysics"
• Code builds without errors: dotnet build OpenFairway.sln
• Tested in Godot editor (if applicable)
• Manual testing performed (describe in comments below)

Code Quality

• Code follows project conventions (PascalCase for public, _camelCase for private)
• Added/updated XML documentation comments for public APIs
• No compiler warnings introduced
• Code formatted with dotnet format (will be checked by CI)

Additional Notes

Related Issues

Closes #