Macbeth is one of those impossible plays, and partly because it’s so simple. Garry Wills has argued fairly convincingly that this shortest of tragedies has in fact acquired its reputation as a cursed play not because more things tend to go wrong in the theater when it’s being performed (statistically, they don’t), but because the performing of it is risky as a theatrical enterprise. It’s a very easy play to mangle, and a bad performance of it is pretty dreadful.
Probably the chief reason, Wills suggests, is that those doing the play don’t trust it. They tend to want to modify it somehow — to make it into something other than what it is. Most often they want to eliminate the supernatural element, which tends to embarrass modern audiences. When one does this, however, it becomes a purely psychological play, of a sort that Shakespeare never wrote (no, not even Hamlet), and it loses its hard objective edge. Banquo’s ghost has to exist for real; the witches have to be making prophecies for real; Macbeth himself has to wrestle with the dual concepts of fate and moral responsibility (never completely happy together).
When one trusts the play, it has remarkable power. Its language is as concentrated and loaded as anything Shakespeare ever produced. It also trades heavily on the whole notion of ambiguity — something that is reflected in the nuances of language at every turn, which is encapsulated in the porter’s ribald but trenchant speech about equivocation (a matter of contemporary political import then as now), and which echoes, as well, the relationship of destiny and free will.
Accordingly, to my taste, the best productions of this play are the ones that have the least invested in putting a topical “high concept” spin on the story. The story is there waiting to be discovered, not covered up. None of them really manages to capture the subtle grandeur of the play. Of the cinematic treatments, the most lavishly faithful is probably the Polanski version, but it is taking some bloody side-trips into the bargain.