The plot skeleton mentioned in the last essay can be linear, a "simple" series of cause/effect alternations in a straight line ("The Little Engine That Could"--the train is on a track with only one task--get over the mountain), or extremely complex, with detours, reverses, braided lines of plot, etc. Every complexity requires some wordage (more wordage from the less efficient writers, less wordage from the more efficient.) Thus the level of complexity is inherently limited by the length of the story, and also by the efficiency of the writer.
In the mystery genre, the relationship between character complexity, plot complexity, and wordage shows up well. Mysteries depend on plot twists--on some complexity of plot--and many readers tolerate less complex secondary characters (especially among suspects, detective's sidekicks/associates) if the plot is convoluted enough. In some mystery lines, wordage is tightly controlled by the publisher; in others, it's not, especially for bestselling writers. Long mysteries (by writers such as Sayers, George, James) spend considerable wordage on the complexity of main (and sometimes even secondary) characters, giving them extra depth so that additional layers of motivation can work in the plot. This can become overpowering to the point of damaging the mystery aspect of the story, if the reader is dragged through too much psychological angst when what he/she wants is to find out whodunnit and see the detective triumph again. Shorter mysteries (60-80,000 words, the typical thinner paperback) provide just enough character complexity in the detective to make him/her interesting, and fill the rest of their wordage with plot complexity...and 90% or more of the plot complexity is directed at solving the crime, not untangling the detective's many other problems.
What this means for the writer is that early on the writer must decide how big a canvas he/she is working on. In a multi-volume story arc, there's ample room for both character complexity (multiple characters) and plot complexity deriving from both outside events and the characters' internal motivations. In a short story, there's not. Writers have a natural bent towards a length, a length that's natural for them, and thus tend to develop characters and plots that fit their natural length. When a natural short-story writer tries to move to a longer length, he/she needs to know that this is going to require not just more events, but more complexity, more connections between characters (who need to be more complex) and plot. The natural novelist faced with a short story assignment needs to know that compression and simplification of both character and plot will need to learn the reverse--how to compress character and plot and not allow rampant growth of complexity.
Another difference between shorter forms and longer forms is the "hinge" or "keystone." Every story--even a short-short--has a place where the story changes direction. In short stories, the hinge may not be noticed by most readers , especially if the story is fast-paced. Even in the most linear, obvious plot, has a hinge (in "The Little Engine That Could", it's where the little engine really commits to the task of dragging the train over the mountain.) Hinges in short forms aren't usually a problem to either writer or reader. Hinges in very long form stories (multi-volume) are often perceived by reviewers (and some readers) as inferior books in the middle of a series....books that do not stand alone, books whose story arc is, by itself less satisfying. The "hinge" phenomenon is seen as a failure only because it's not understood. In a good multi-volume work, the story arc spans the entire group of books: the group comes to a real climax, a true ending. Once the group is complete and can be read as a whole, that hinge volume is no more a failure than the middle section of a one-volume novel, or the middle paragraph of a short story, is a failure. It's a necessary change of direction, part of the essential nature of Plot, and thus inevitable. (It's also the reason odd-number groups read better than even-number groups...in odd-number groups, the middle volume will be the hinge and the least stand-alone; in even number groups, two volumes--one either side of the middle--will be affected. Writers whose natural length is very long should consider thinking in 3, 5, or 7 volume units, not 2, 4, 6.)
I've decided to break up the Plot discussion, because covering the many ways to make Plot more complex and satisfying without violating its requirements is going to take (me, anyway) a lot of words.