driver_isolation_identity_terms

The cumulative-identity split of tier 3, aggregated over any group of fct_driver_isolation_pair_lap rows (WI-16a).

The cumulative-identity split of tier 3, aggregated over any group of fct_driver_isolation_pair_lap rows (WI-16a). Use inside a GROUP BY over that model. Per pair row, relative = pace_gap + car_advantage + traffic_advantage + tyre-age bias, where relative is the raw gap (W63 design 2, 2026-10-03: no tyre-age adjustment, peers matched within var(‘isolation_peer_age_tolerance’) laps of tyre age instead) and the bias is the compound seed’s pricing of that age gap, (C(peer) - C(driver)), which pace_gap carries and the raw gap does not. It is 0 between cars on the same tyre age and is not published. Every term is a CONTRIBUTION: its sum over the group’s closed rows (both cars have a car term, so pace_gap is non-NULL) divided by identity_n_pair_laps. Then identity_relative_pace_gain_s = identity_pace_gap_gain_s + identity_car_advantage_gain_s + identity_traffic_advantage_gain_s + (the group’s mean tyre-age bias) The three published contributions therefore close identity_relative_pace_gain_s only up to that bias. T42 asserts the three-term identity exactly on same-age pair rows, where the bias is 0. A contribution over zero rows is 0, not a measurement. All columns are NULL when the group has no closed row. Seconds per lap, positive = in the focal driver’s favour. Emits: n_pair_laps, identity_n_pair_laps, identity_relative_pace_gain_s, identity_pace_gap_gain_s, identity_car_advantage_gain_s, identity_traffic_advantage_gain_s.

Signature

{{ driver_isolation_identity_terms() }}

Usage

SELECT stint_id, {{ driver_isolation_identity_terms() }}
    FROM {{ ref('fct_driver_isolation_pair_lap') }}
    GROUP BY stint_id