Analogue and digital signals, the binary number system, analogue-to-digital and digital-to-analogue conversion, resolution and quantisation.

An analogue signal varies continuously and can take any value within a range (the voltage from a thermistor divider, which changes smoothly with temperature). A digital signal has only discrete levels (in a binary system, two: logic 0 and 1), and information is represented as binary numbers.

A microcontroller works in binary and can only process digital values, but a temperature sensor gives a continuously varying analogue voltage. An analogue-to-digital converter (ADC) is therefore needed to sample the analogue voltage and convert it into a binary number the microcontroller can read and process.

An 8-bit ADC provides $2^8 = 256$ discrete levels (codes 0 to 255).

Markers reward the continuous (analogue) vs discrete (digital) contrast, the reason an ADC is needed (the microcontroller is digital but the sensor is analogue), and $2^8 = 256$ levels.

The number of steps for an $n$-bit ADC is $2^n$, and the resolution (size of one step) is the range divided by the number of steps. (Using $2^n$ steps.)

For a 10-bit ADC over 0 to 5 V:

So each step represents about 4.9 mV.

If more bits are used, there are more steps ($2^n$ grows), so each step is smaller and the resolution is finer (the digital value follows the analogue signal more closely, with less quantisation error).

Markers want $2^{10} = 1024$, the resolution of about 4.9 mV (5 V / 1024), and the statement that more bits give finer resolution.