I have seen and read several videos and articles that try to explain how Zero Knowledge Proof systems work. In these examples, there is a verifier of information, and a prover of information. The verifier asks the prover to prove that they know some information. Metaphors I have seen usually are:
The prover wants to prove they are not colorblind and gives the verifier two colored items. The verifier is color blind and switches the items around (or not). The prover can tell if the items were switched. Repeat this a hundred times successfully and the prover has proven to the verifier that they really are not colorblind.
I've seen variations with Pepsi/Coke taste tests, colored pens or balls, etc. However, this kind of proof requires that the verifier has some sort of information at their disposal already (whether they switched the balls around or not).
I've also heard examples such as:
A liquor store asks a customer if they are at least of drinking age. The customer can prove they are without having to show their ID.
A bank asks a customer to prove they have a minimum balance so that they can get a loan, without having access to the customer's bank account.
These are examples where the verifier (the liquor store and the bank) doesn't have the same information at their disposal (such as the example of switching the items around).
If a liquor store asks a customer if they're of drinking age, what's preventing the customer from just lying as many times as needed to buy their alcohol?