int_era_normalized_driver_rating

One row per (driver_id, season).

Part of the Skill family.

One row per (driver_id, season). Cross-era driver rating: compares drivers across 2018–2024 despite regulation changes (2022 ground-effect boundary). Uses bridge drivers (≥8 races both eras) to estimate era offset, then adjusts season residuals. era_adjusted_rating = comparable across eras; negative = faster than era average. low_anchor_sample_flag = few bridge drivers (weak offset).

Lineage

Overview

Property	Value
Layer	intermediate
Family	Skill
Contract enforced	No
Tags	driver_rating, era_rating
Upstream	int_driver_season_ratings

16 tests guard this model - 15 generic, 1 singular.

Columns

Column	Type	Nullable	Tests	Description
driver_season_id		Yes	not_null, unique	Surrogate PK = hash(driver_id, season). Matches int_driver_season_ratings.
driver_id		Yes	not_null	Driver identifier
season		Yes	not_null	Race year (integer)
raw_residual_mean_s		Yes	not_null	Unadjusted season mean. Not era-adjusted.
shrunk_residual_s		Yes	not_null	Bayesian posterior mean (season shrinkage applied, no era offset).
era_adjusted_rating		Yes	not_null	Cross-era comparable rating. Negative = faster than era average. Key output column.
era_adjusted_rating_se_s		Yes	not_null, expect_column_values_to_be_between	Propagated SE (sqrt of shrunk SE² + era shift SE²).
era_adjusted_rating_ci_low_s		Yes	not_null	Lower bound of 95% CI on era_adjusted_rating
era_adjusted_rating_ci_high_s		Yes	not_null	Upper bound of 95% CI on era_adjusted_rating
rating_confidence		Yes	not_null, expect_column_values_to_be_between	n_races / (n_races + 5). [0, 1].
bridge_driver_anchor_flag		Yes	not_null	TRUE if this driver was used to estimate the era offset.
low_anchor_sample_flag		Yes	not_null	TRUE if fewer than 3 bridge drivers found; era offset is 0 when this fires.