The gradualness of mechanical failure lets it affect operation before it is noticed

What it is

Missed keys rise, people slam to bottom, a knob walks in the car; users often change their own motion to cope rather than file a repair. Gradual faults degrade use before they are noticed. Feel fade is what happens physically. This layer is the human side: adaptation hides the fault, and accidents and efficiency losses happen before anyone says “the key is broken.”

Why it happens

People adapt to slow change: a little more force, a glance at the screen, less eyes-free use. Those compensations keep the primary task completable, so there is no crisp failure event to report. Compensation itself has a cost—gaze tax, false hits, not daring to let go in vibration—filed as “it’s been a bit awkward lately” or “I’m clumsy.” Only when compensation is not enough (slamming never lights, or walk drives a value into danger) does it enter service. On a safety-critical eyes-free channel, adaptation looks like “started looking at the device,” which is exactly what the channel was meant to avoid. Field data that only counts replaced parts systematically undercounts the stretch of life already harming use.

Studying it

Follow the same users or the same devices longitudinally, and at points before hard failure run eyes-free tasks and score compensation (looking, slamming, two hands).

Independent variables: cycle progress, whether a new-part feel reference is offered, whether compensation is asked about. Dependent measures: eyes-free hits, compensation rate, the point of voluntary repair, primary-task score.

A one-shot end teardown will not see adaptation. Quietly swapping in a new part and watching scores jump can show that fade was being hidden. Do not only ask “is this key fine”; adapted subjectivity will say yes.

Where it stops holding

A sudden break (it shatters on one press) has no such latent stretch; the gradual claim does not apply. Professional shops with scheduled calibration catch feel-escape before users adapt; consumer electronics rarely have that calibration. When users replace devices often, fade is not yet adapted to and the device is already gone; the issue becomes “it doesn’t feel like it used to” across brands. A safety audit that only looks at hard fails will miss the cohort already operating by looking.

Applying it

• Treat compensatory behavior (slamming, looking, giving up eyes-free) as a fault signal in support and telemetry; do not wait for an open circuit.
• For eyes-free safety keys, specify in-field spot checks that they must still complete eyes-covered, rather than waiting for a ticket.
• Verify: at mid-life, an undeclared eyes-free test against a new part. A drop in hits or a rise in looking is already latent failure. In support scripts, collect sentences like “I have to slam it,” and file them as feel-escape, not as user error.

Related

• Same group: C10.16.1 A physical control’s life is usually measured in presses or turns it can survive · C10.16.2 Mechanical wear changes key travel and damping feel gradually, rather than failing all at once · C10.16.3 High-use controls need a higher rated life than low-use ones
• Adjacent: C10.08 Key Travel, Tactile Detent, and Confirmation · C10.09 Eyes-Free Key Layout Zoning
• Search: silent degradation · compensatory behavior · undetected wear