For context, in 1929 the priest and detective author Ronald Knox "codified" a set of rules that, in his opinion, divided low- from high-quality detective stories, and ought to be followed as much as possible in order to create a "fair" playing field between the author and the reader. They are:
1. The criminal must be mentioned in the early part of the story, but must not be anyone whose thoughts the reader has been allowed to know.
2. All supernatural or preternatural agencies are ruled out as a matter of course.
3. Not more than one secret room or passage is allowable.
4. No hitherto undiscovered poisons may be used, nor any appliance which will need a long scientific explanation at the end.
5. No Chinaman must figure in the story. (This one is kind of the elephant in the room. It doesn't come across very well now, but it was a reaction to the overuse of "exotic" elements in detective stories at the time, when they were added for the sake of melodrama without contributing anything to the mystery itself. I'm not going to name the book, but there was a mystery novel published recently, which featured several well-developed "foreign" characters, and one who seemed like he was only there to clean up after the rest of the cast, keep their secrets, and do the dirty work in bringing down the bad guy. But, unlike in Knox's day, you couldn't immediately peg him as the bad guy.)
6. No accident must ever help the detective, nor must he ever have an unaccountable intuition which proves to be right.
7. The detective himself must not commit the crime. (Spoilsport.)
8. The detective is bound to declare any clues which he may discover. (Personally, I think the detective should be allowed to hide one clue per book from the reader, but that the narration should provide some hint of what it is.)
9. The "sidekick" of the detective, the Watson, must not conceal from the reader any thoughts which pass through his mind: his intelligence must be slightly, but very slightly, below that of the average reader.
10. Twin brothers, and doubles generally, must not appear unless we have been duly prepared for them.
Now — obviously these rules were often disregarded, even in Knox's day; Arthur Conan Doyle and Agatha Christie broke them all the time, and are still widely considered the co-regents of the genre. (The idea for this post came from a discussion of "The Murder of Roger Ackroyd" a day or two ago.) In addition, pure whodunit stories generally seem to be a lot less common now than they were in the early 20th century — but I would say that it's still a popular genre, and one that deserves to be analyzed in modern terms, rather than treated only through a historical frame.
All of which is to say, which of these rules are "keepers"? Which ones can be adapted to the modern form of the genre with a little effort, and which ones need to be thrown out?
by Raineythereader