Lisp's Mysterious Tuple Problem

Even with accessors like (point-z p), a structure is not more verbose than (caddr p) or (third p); it can't be that someone avoids using a struct because (third p) is much less verbose.

I don't know of a Lisp programmer who wouldn't write a function called point-z (or whatever) to make it clear that it accesses a Z coordinate, even if a list is used during prototyping, so that the code is more readable and easier to change later.

Most Lisp programs and programmers avoid structures? Where are the numbers on this? Maybe the author looked mostly at Emacs Lisp code bases.