I suppose it depends on the specific application. If I were talking to emergency responders and they happened to mishear part of the message, I'd much prefer the chance that they have an idea of my vicinity than having a likely infeasible number of possible word combinations, a few of which might provide reasonable locations if it were possible to check them all.

In other words, your example would hold true if they heard and remembered the three words correctly in the wrong order, but that's not a gamble I'd want to take particularly if the API to check codes is restricted.