Inconvenience or Impairment? A Fixed-Criterion Protocol for Measuring What Habitual Full Self-Driving Users Supply When the System is Unavailable

Every supervised self-driving trip can end with a person back in charge of the car. No public safety number isolates that person. Collision rates for Full Self-Driving (Supervised) describe the driver and the system together, as used; they do not show what the driver can still supply when the system is unavailable. Reserve is the distance between what you can do and the fixed tasks of an independent life. Help that does what you could still do can spend it. Driving is one of those tasks, and for older adults one of the hardest to give up. This paper proposes the measurement: a handback challenge. Every participant is tested at enrollment, before habitual use, and at six and twelve months, beside owners who use the system little or not at all. Hazard response is the primary safety outcome, judged against the pass mark of a licensing test whose failure predicted crashes in new drivers. Vehicle control is checked with the standardized on-road highway test, whose weaving measure carries an established margin equivalent to a 0.05% blood alcohol concentration. A simulator takeover after twenty minutes of identical simulated automation tests the moment control returns, and a block with an added task tests spare capacity. A decision table assigns each result to no measurable cost, inconvenience, spent reserve, or transient, reversible or durable impairment. Aviation and navigation research point to a dissociation, which this paper predicts for driving: hand skills kept, hazard skills lost. The vehicle-control check alone needs 36 to 78 drivers per group. Three analyses of data the manufacturer already holds, one of them built on a per-vehicle score it computes today, would test the same question across the fleet.