int_tyre_surface_vs_bulk_decoupling

One row per lap (post-cliff only).

Part of the Residual family.

One row per lap (post-cliff only). After tyre cliff, diagnoses whether degradation is surface-recoverable or bulk-structural. surface_bulk_ratio [0, 0.606], 0.5 at steady push: high = recoverable, low = must pit. Includes recovery_probability (pace-recovery chance within 2 laps). Grain: lap_id filtered to cliff_onset_passed = TRUE.

​

Lineage

​

Overview

Property	Value
Layer	intermediate
Family	Residual
Contract enforced	No
Tags	simulation, thermal
Upstream	int_lap_thermal_proxy, int_lap_residual_decomposed

10 tests guard this model - 9 generic, 1 singular.

​

Columns

Column	Type	Nullable	Tests	Description
lap_id		Yes	not_null, unique	PK, FK to stg_laps (post-cliff laps only)
surface_bulk_ratio		Yes	not_null, expect_column_values_to_be_between	Surface share of the two thermal loads, each divided by its own weight sum (vars thermal_surface_weight_sum / thermal_bulk_weight_sum): 0.5 at steady push. F49 (WI-15b): surface <= bulk by construction, so the ceiling is bulk_sum/(surface_sum+bulk_sum) = 0.606 (all push on the current lap); the bound below is that ceiling, not [0, 1]. Before F49 the unnormalised share was capped at 0.5.
degradation_source		Yes	not_null, accepted_values	bulk_driven when surface_bulk_ratio < 0.35, else mixed. F49 (WI-15b) removed surface_driven (> 0.65), which the ratio’s 0.606 ceiling made unreachable. T39 (assert_degradation_source_classes_reachable) checks every value below has rows.
recovery_probability		Yes	not_null, expect_column_values_to_be_between	Probability of pace recovery within 2 laps. [0, 1]. SQL approximation.
thermal_attribution_s		Yes	not_null	Share of compound_component_s attributable to thermal (surface) damage.