"Now, naturally, I've never written a mystery novel myself, but I've read quite a few and I know what kind of solutions the reader likes and the kind he doesn't like.
And, so, I suggest the following rules when writing locked room mysteries:
* Claire's 1st: The key (or keys, if by some bizarre chain of events) found inside the locked room must not be planted on the corpse (or somewhere else in the room) after or during the body's discovery - be it by the culprit or an accomplice.
Honestly, at this point - it's a cheap trick and obvious. It's the first thing the reader will think while the locked room is being established. The simple fact of the matter is that, with a solution like this, a locked room doesn't exist, at all. The very promise of an impossible and daring crime is a flat-out lie. Usually, this is combated by having the detective discover the key to the room themselves and prove to the reader that it was really there from the beginning OR add an element to the locked room which proves that it's very unlikely the culprit ever locked the door from the outside (such as a chain that can only be set from the inside).
* Claire's 2nd: All the entrances and exits to the room should be made absolutely clear to the reader. In addition, it should be made absolutely clear how exactly they work.
This is to eliminate secret passages and such, but also things like a hidden auto-lock or something of the sort. Having a solution that revolves around the reader's lack of information is weak, really.
* Claire's 3rd: No mechanical contraption (ie, trap) that would allow the culprit to kill without being inside of the actual room should be used to kill the victim.
This one is, admittedly, more of a personal preference, but I don't like the idea that the crime should boil down to 'how did this trap that killed the victim work' instead of 'how did this murder work'. Not everyone is an engineer or interested in mechanical contraptions or such. The puzzle then becomes nothing more than just arranging the pieces on the crime scene in a way that makes it possible for the trigger to be pulled. And those become more bothersome than interesting. There is no chase. Mysteries that have to do things like this also often give all the suspects a perfect alibi or something and make the thing all the more obvious.
Naturally, the nature of the trap itself or the reason it was used in the first place often leads to the actual culprit, but... it no longer feels like a true impossible crime when a reader can feel just as satisfied going 'A trap was used!" and call it a day (which is, I imagine, most people would do instead of figuring out how the trap worked, anyway).
* Claire 4th: When having corporate alibis (ie, alibis in which the suspects themselves provide each other air-tight alibis), at least one of the suspects should be completely innocent in the crime itself, speak the absolute truth and have no reason to hide anything in regards to that alibi. (As characters, they can still have secrets - but it must be set in stone that the alibi provided is valid!)
This is mainly to help establish trust between the reader and the author. After all, if everything can be explained with something as simple as 'they all lied', then we get into the territory of third-rate in which Markham's novel had found itself in.
* Claire's 5th: Regardless of who the narrator might be, they must provide the events in an objective way and must not leave out any relevant details that would be obvious to anyone that might've been with them at the time. If the narrator themselves is the culprit (assuming that the narrator is in first-person, a third-person narrator being the culprit is beyond ludicrous), then it is allowed for them to omit details in which the nature of the trick is revealed (ie, the moment of the actual crime and how it was done), but it should done be in a way where the nature of the problem is set before the reader properly and explained in full just as the clues to the mystery are still presented. In addition, the narration must never flat-out lie to the reader when it comes to the events (in the case of the culprit-narrator, the narrator's thoughts must not conflict with the actual solution or be misdirecting to the point where they don't make any sense on a second read).
This one is pretty self-explanatory. It's basically the equivalent of Knox's 8th: All the clues to solve the mystery must be presented. The only addition to this is the fact that the narration must not lie, which is obvious enough. This is more of a literary rule than an actual locked room one and can pretty much apply to mysteries in general, but it's essential to establish even in crimes like these, as some people seem to think that petty deception like this makes their locked rooms automatically brilliant.
A rule like this, as you can see, does not require 'a detective' (somebody who had an undeniable, objective view) like Knox or Van Dine do, but it still follows the notion of objectivity. I am prepared to give enough leeway that, if some people wish to do so, they could make their narrator secretly a culprit. They'll only get to do it once in their careers, anyway.
* Claire's 6th: The culprit, accomplice, or anyone else involved in the crime itself must be people that have not only been mentioned at the story, but whose physical presence among other characters has been established (ie, they've interacted with other characters in some form -- besides killing them, that is).
This is another literary one, admittedly.
What I mean by this, essentially, is that it's not enough to simply say 'this person is relevant, but you never got to meet them, but they are one of the people responsible in pulling this off'. No. The reader must have met them and the nature of their character sometime before the murder and they must have somehow interacted with other characters at said point (ideally, the detective). It's the equivalent of Knox's 1st, if you will - but a bit more restrictive. It's to prevent the writer to pull a rabbit out of their hat at the last moment and make a getaway, claiming that we simply never considered the right person as a suspect, blinding us to the fact that we were never informed we should ever do such a thing to begin with.
This rule also eliminates the possibility of twins or look-alikes. Those are still considered to be different characters, with different personalities and, evidently, different agendas. The only way I could see this working is if it was properly hinted at, but doing so would probably end up with the reader catching on to your scheme long before the actual crime takes place, so... be warned.
Naturally, if you have a mystery set with people trapped on a secluded island or something of the sort, a hidden person X is out of the question, obviously.
* Claire's 7th: There can only exist one solution.
You may find this one to be another personal rule, but it makes sense, in my opinion. Naturally, you are allowed to have characters in the novel speculate about the locked room and reach their own conclusions - it'd be strange if they didn't. However, every other theory aside from the solution should be deniable either by common sense or evidence found at the scene of the crime (or it simply doesn't happen to explain all the evidence found on the scene). There should be a path that supports only one possible conclusion that leads to a single answer to the whodunnit and howdunnit. It is permitted to have different possible explanations for the whydunnit, but the final explanation provided should be the one that makes most sense and the one that was hinted and foreshadowed the most in the actual book.
... What do you think?"
Neither Markham or I could quite think of how to respond to all of that for a little while.
Finally, he let out a bit of a nervous chuckle before elegantly wrapping up the whole matter regarding the manuscript idea and the discussion on mystery novels in general. "You should be a critic, not a detective."
"If there's one thing that could possibly qualify me as being a critic - it's that I'm a detective. And I'd much rather be a detective than a critic. After all, these rules apply to fiction, not the real world. The real world is much less ideal in the sense that the criminals are far more incompetent and unorganized. Which, naturally, makes my - and the police's - job a lot easier. Whether that's a good thing or not - I'm not quite sure.
After all, I alw
ays did enjoy a challenge..."
###
Now then - I must admit that this second story of mine (after The Problem of the Straton Murder) isn't really that much of a short story so much as it is my personal view on the mystery genre - well, specifically, locked rooms. I do not believe I need to elaborate further than what has been said in the actual story. The rules stated here are the ones I intend to follow myself through my future... 'attempts', I suppose you could call them.
Whether following any of them will lead me to success or not in the formation of those future 'attempts', we shall have to see, admittedly.
-- J. R.
The Rules of the Game Page 2