Lists of pontifical dicta have been drawn up by various experts, but they seem to me to be a lot of nonsense....The most frequently repeated rule, generally assented to, is the most nonsensical. It says, "You must play fair with the reader," meaning that in the course of the narrative the reader must see and hear everything that the detective sees and hears. I don't know why people like S. S. Van Dine and R. Austin Freeman and Dorothy Sayers have insisted on it, since every good writer of detective stories, including them, has violated it over and over again.
Stout says Sherlock Holmes often found evidence he did not tell Watson (and the reader) about, and it "is the same with Dupin, Lecoq, Father Brown, Poirot, Wimsey, Perry Mason--practically all of them." He declares that a detective fiction writer need not worry about practicing "fair play," then, but s/he does need to have "the required qualifications for a storyteller":
You can create people in your mind that you get excited about; you can devise and develop a plot situation, you have a sense of structure and form; and you can write readable narrative and dialogue. Thus equipped, you can write good stories, but it doesn't follow that you can write good detective stories. For them you need something more.
What, you must be asking. Stout answers:
You need, not the kind of mind that likes to solve puzzles, but the kind that likes to construct puzzles, which is quite different. You also need considerable ingenuity if they are to be not only puzzles but good ones.
What do you think? Is "fairness" in the presentation of clues important to you in a classical detective fiction, or do you care more about the other elements--storytelling skill and ingenuity in puzzle construction--outlined by Rex Stout?