Perhaps it is trivial, but what's the point? If you can reason what algo to use and reason how to determine which one will be the most efficient of the two without testing them both (or explain the abstraction which allows you to test them both without significant overhead in code), it should be enough.
That doesnt follow to me. In theory, there is a difderence between theory and practice, because theory simplifies reality into a model with given assumotions, and reality...is reality.
