I disagree with your example being a bad and reason is right there in your comment: the word "protocol". From wikipedia we find that a "communication protocol is a system of rules that allow _two or more entities_ of a communications system to transmit information...", where I have underscored the essential phrase and cryptographic protocols are just a subset of communication protocols.
The reason that Alice and Bob are introduced is because they are different entities and each has different powers, restrictions and goals. If you want to do a correctness or security analysis of a protocol, introducing and defining these parties is hugely important and useful. Otherwise, its very easy to make mistakes about where information is in the system and how it is being manipulated. More generally, there is a philosophical approach called operationalism [1] that demands that all science theories be couched in terms of protocols so we don't make mistakes about what they say.
(A)lice and (B)ob set up a channel.
(E)ve is an evesdropper.
When the only thing you change is to personify A, B, and E in your narrative description of a logical protocol, you are not making an analogy or an abstraction.
Sequence diagrams, BAN logic, UML, pseudo code and other protocol notations are not teaching analogies, they are specifications. Some more useful than others.
Teaching analogies and abstractions are more of a mix of use cases, literally analogous functional relationships, black boxing, edge cases and boundary conditions. The "why" behind the "how."
That Alice and Bob can be replaced in every situation with (A) and (B) means they are not abstractions or analogies, and are therefore patronizing cartoons that obfuscate useful information.
Alice and Bob aren't just A and B. A and B are members of the category of "things"
Alice and Bob are members of the category of "communicating entities", and are also familiar names for human. Humans communicate, so choosing Alice and Bob makes an analogy that helps the learner remember that A and B are the communicating entities, not messages, or the encryption key, or anything else.
They are illustrative examples that help with visualization and connecting the theory to an interesting use-case, which taps into most humans' natural preference visualization, socialization, and relevance. They aren't patronizing, they are humanizing.
Do you also feel that speaking in English is a "patronizing" way to communicate, since everything you say could be expressed in mathematical symbols?
That's a more complex idea that requires a more complex model. For example, Alice and Bob change to mailboxes, and Alice can have multiple mailboxes and mail stuff to herself.
The reason that Alice and Bob are introduced is because they are different entities and each has different powers, restrictions and goals. If you want to do a correctness or security analysis of a protocol, introducing and defining these parties is hugely important and useful. Otherwise, its very easy to make mistakes about where information is in the system and how it is being manipulated. More generally, there is a philosophical approach called operationalism [1] that demands that all science theories be couched in terms of protocols so we don't make mistakes about what they say.
[1] https://plato.stanford.edu/entries/operationalism/