Thanks for helping me to understand I (partly) understood 🙂

Ok, given such encoding of discrete values in a single qubit will always have a certain uncertainty, how do you deal with?

In your example it is quite obvious because you only have a single operation in your circuit. So you set the expected distribution up, run the circuit and measure (sample) outcomes. In other words, in this case you can simply compare observed to expected.

But what if the circuit is more complex, consists of multiple operations resulting in a unknown distribution of outcomes. In other words, you cannot directly compare observed to theoretical expected, so you cannot detect ´encoding errors´.

Would it even matter to know there are errors?
Or am I now drifting into ´missed the point´ territory?