When the reading is wrong

Five symptoms

Pick what the reading is doing. The ringed part of the board is where to look first, and the list beside it is the order to check in.

Stuck at 0

SIGNAL is at 0 V wherever the wheel is. The usual reason is VCC: with nothing on it the track has no top, and the wiper reaches GND through it. If VCC is connected, check that POT_PIN names the pin SIGNAL is really on, and that the cable is not one pin over. Shifted one position, SIGNAL lands on NC, which is connected to nothing.

Stuck at full scale

SIGNAL is at VCC wherever the wheel is. The usual reason is GND: without it the track has no bottom, and the wiper sees VCC through it.

In MicroPython on an ESP32, a missing atten() line gives a reading that hits the top under a third of the way round, because the ADC's default range there is about 1 V. In Arduino, an ADC_MAX of 1023 on an ESP32 makes a percentage or a duty hit the top a quarter of the way round.

Wanders on its own

A reading that drifts over hundreds of counts with nobody touching the wheel means the pin is measuring nothing. SIGNAL is not reaching it, and an unconnected analog input reads whatever charge happens to be on it. Check the SIGNAL wire, then the pin number.

A few counts of jitter with everything connected is normal, more on an ESP32 than on an Uno. Smoothing a jittery reading is the fix.

Goes the wrong way

Which direction raises the reading is set by the part, and nothing is wrong. Use ADC_MAX - reading instead of reading, and the wheel works the other way round.

Stops short of the ends

On an ESP32 the last few percent at each end read flat, and on any board a stop may be a few counts short of 0 or full scale. Why the ends are not exact has the sketch that takes your own stops.