S.S Van Dine's Twenty Rules for Writing Detective Stories

S. S. Van Dine (1888 - 1939) was the pseudonym of Willard Huntington Wright.
Van Dine was born in Charlottesville, VA. He worked as a literary and art critic for newspapers and magazines. He suffered from poor health, had a severe breakdown in 1923 and was confined to bed for two years. During these years, he read detective stories and amassed a large collection. He then decided that he could write a better story than he was reading. His first Philo Vance book, The Benson Murder Case, was published in 1926. His books were exceptionally popular, and Van Dine became quite wealthy.

In 1936, S.S. Van Dine (author of the Philo Vance mysteries) published an article titled "Twenty Rules for Writing Detective Stories." Some of his rules are a bit archaic, but it seems to me that they are still worth thinking about.
However even the great have some times flouted these rules (and gotten away with it).

1) The reader should have the same opportunity as the detective to solve the crime.

2) No tricks can be played to mislead the reader unless it is also done to the detective by the criminal.

3) The detective should not have a love interest.

4) Neither the detective nor one of the official investigators can turn out to be the criminal.

5) The villain must be found by logical deduction, not luck, accident, or un-motivated confessions.

6) The story must have a detective who also solves the crime (by detection).

7) It must be a murder mystery ("the deader the corpse the better").

8) The solution must come by "naturalistic means"; e.g., no ouija-boards.

9) There can be only one detective; not a team.

10) The villain has to be someone who plays a prominent part of the story.

11) The culprit can't be a servant (none of this, "The butler did it.").

12) There can only be one murderer. The villain could have a helper or "co-plotter," but only one is going to get the ax in the matter.

13) No secret societies ("mafias, et al"). The murderer, too, needs a sporting chance to outwit the detective.

14) The method of the murder must not be beyond plausibility. No super-natural means, nor the introduction of a fictional device or element ("super-radium, let us say" is not fair).

15) The truth of the solution must be apparent. The reader should be able to pick the book upon completion and see that the answer was in fact starring at him all the time.

16) The detective "novel" must be just that, no side issues of "literary dallying" or "atmospheric preoccupations." These devices interfere with the purpose of detective fiction, "which is to state a problem, analyze it" and solve it.

17) The culprit must be an amateur, not a professional criminal.

18) The solution must never be an accident or suicide.

19) Motives for the crime must be personal, not political or professional.

20) All of the following tricks and devices are verboten. They've been done to death or are otherwise unfair.

a) Comparing a cigarette butt with the suspect's cigarette.
b) Using a séance to frighten the culprit into revealing himself.
c) Using phony fingerprints.
d) Using a dummy figure to establish a false alibi.
e) Learning that the culprit was familiar because the dog didn't bark.
f) Having "the twin" do it.
g) Using knock-out drops.
h) If the murder is in a locked room, it has to be done before the police have actually broken in.
i) Using a word-association test for guilt.
j) Having the solution in a coded message that takes the detective until the end of book to figure out.

References
Mystery Novels of the Golden Age
http://mikegrost.com/vandine.htm
http://www.mysterynet.com/books/testimony/forgottenmaster.shtml
S.S. Van Dine's 20 Rules for Writing Detective Stories

PD James:

What is the difference between the detective story and the crime novel?

I see the detective story as a subspecies of the crime novel. The crime novel can include a remarkable variety of works from the cosy certainties of Agatha Christie, through Anthony Trollope and Graham Greene, to the great Russians. The detective story may be considered more limited in scope and potential. The reader can expect to find a central mysterious death, a closed circle of suspects each with credible motive, means and opportunity for the crime, a detective, either amateur or professional, who comes in like an avenging deity to solve it, and a solution at the end of the book which the reader should be able to arrive at by logical deduction from clues presented by the writer with deceptive cunning but essential fairness. What interests me is the extraordinary variety of talents which this so-called formula is able to accommodate.