Mysteries do not need to be solvable

In the past year I’ve gotten into reading mystery novels, and this has reinforced one of my strongly held opinions about the genre. There is a mistaken preconception about mystery novels, that the reader ought to be able to solve the mystery. This simply is not true. There are some mystery stories that are meant be solvable, but it’s a minority of mystery stories that I’ve seen. Solvability is not the primary appeal of the genre, or at least it’s not the appeal to me.

The reason I know this, is because when I was young, we had a “complete works of Sherlock Holmes” book, which had all the short stories. I didn’t read them all, but I read enough to know that Sherlock Holmes stories were not solvable. Usually, Sherlock Holmes would pull some clue out of thin air, that hadn’t been mentioned before; or else there would be an event that led to the mystery being solved. It was unambiguous that most stories were not even trying to be solvable. The mysteries were trying, first and foremost, to be stories. There’s something the reader doesn’t know (rising tension), and then Sherlock Holmes explains it (releasing tension), and that’s a simple but effective narrative arc.

Golden Age Mysteries

The idea that mysteries ought to be solvable is a later idea, strongly promoted during the “Golden Age of Detective Fiction”. A brief timeline: Edgar Allan Poe is often credited as the originator of English detective fiction, with Auguste Dupin in 1841. Arthur Conan Doyle popularized the genre with Sherlock Holmes who appeared from 1887 to 1927. The Golden Age of Detective Fiction was in the 1920s and 1930s, and the most famous author from that period is Agatha Christie. The Golden Age declined in popularity in World War II, and the genre we have now has greatly branched out, to the “hardboiled detective” subgenre, to procedurals, to crime thrillers, and so on. The current state of the genre is that some mysteries try to be solvable, harkening back to the values of Golden Age mysteries, but this is merely one of many directions that a mystery can take.

In order for a mystery story to be solvable, they need to obey certain conventions, agreed upon by both authors and readers. The most famous codification of these conventions are Ronald Knox’s Ten Commandments, from 1929:

  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.
  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.
  8. The detective is bound to declare any clues which he may discover.
  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.

To my eyes, Knox’s commandments have a certain specificity, as if to express grievance with stories that Knox found unsatisfying. Modern readers tend to ogle at rule #5, and my understanding is that it’s a rejection of a particular racist Chinese stereotype that isn’t nearly as common anymore. The rules contain some very constraining assumptions, such as the idea that there is a detective and a Watson. And then there’s stuff that just doesn’t make sense: if secret rooms and passages are problematic for a solvable mystery, then why is even one allowed?

Knox’s rules came to my attention through Umineko When they Cry, a visual novel that explicitly describes its own version of the Knox Decalogue. It’s interesting to compare to Knox’s original rules. Umineko drops the Chinaman rule, completely forbids hidden passages, and #10 talks about disguises rather than twins. #9 turns into a completely different rule about how eyewitnesses are permitted to offer their own (possibly inaccurate) interpretations.

I think this illustrates that Knox’s rules are useful in some form, but readers and authors may disagree on the specifics. And that creates a problem. What if I believe in a rule that completely forbids hidden passages, but I read a book by an author who believes one hidden passage is permitted? I might feel unsatisfied or even betrayed when the author reveals a solution involving a hidden passage. To me, the mystery was unsolvable, because it violated what I thought were the puzzle’s guardrails. So does the author need to explicitly declare the rules within the story? Umineko gets away with declaring its own rules, because it’s very long and very meta, but I think it would be weird and tiresome if every mystery story did that. In practice, when you go into a “solvable” mystery, you cannot expect it to obey any particular set of rules; all you know is that the author is at least trying to make the mystery solvable.

I think there are two major flaws to the idea of a solvable mystery novel. The first flaw, is that a solvable mystery is in opposition to a natural story arc. If a mystery is solvable in any sort of practical sense, it stands to reason that some readers will solve it. But unless they’re all solving it at the right moment in the story, that takes the wind out of the story’s climax. Therefore, authors are incentivized to make their mysteries very difficult. Golden Age mysteries want to be solvable, but they don’t actually want to be solved, at least not most of the time. The ideal mystery solution is one that makes you say, “Dang, I almost had it!”

The second major flaw, is that both authors and readers can get away with a mystery being unsolved. If you can’t solve a mystery before its solution is revealed, then you just read onwards, and the solution is revealed. If you come up with a solution to a mystery, but it’s not the same one that the author had in mind, then you just read onwards, and the intended solution is revealed (and there are no opportunities for retries). The book still functions even if there was no chance of ever actually solving the mystery. So how can you trust the author to make a solvable mystery? And if you’re the author, how can you trust the reader to make a good effort?

I would analogize the situation to the era of arcade cabinet games. Yes, arcade games were beatable. But they were designed to be very difficult to beat, because that’s who they wrung out more quarters from players. Realistically, they were only beatable by experts, and typical players were left in the dust. Likewise, “solvable” mysteries are often only solvable by long-time fans, who use genre savviness as much as they use detective logic and reasoning.  And how do you become an expert?  If the appeal of a mystery is entirely based on actually solving one, wouldn’t you have to trudge through a lot of mystery stories before you could ever enjoy one?

Alternative Approaches

The mystery genre is myriad, but let me explain what I like about mystery stories. Mystery stories are character-centric. To have a mystery, you need to have suspects, and you need lots of them, or the answer would be obvious. One of the characters needs to be hiding something, but it’s really better if every character is hiding something. This guarantees that each character is, at minimum, complex enough to have two sides to them. Mysteries also encourage attentiveness to the characters, because you just know that some of the minor details will become important later.

My ideal mystery, is one that features a series of wacky suspects. The characters have contrasting and conflicting perspectives, and the detective is somewhat critical of them, but also it’s not entirely their concern because they have a job to do. I do not mind if a mystery tries to be solvable, and no shade on Golden Age mysteries. But I also do not care if a mystery actually succeeds in being solvable. Because to me, the appeal of the mystery is really that it lets me indulge in my desire to critically examine people.

That’s not to say I don’t like a good puzzle. I am a long-time puzzle enthusiast, and it is because I love puzzles that I understand that mystery novels are not a great medium for them. I truly think that video games are a superior medium.

Video games address most of the problems I have with the solvable mystery novel. A video game can block your progress until you solve the mystery. Therefore, the game cannot get away with being unsolvable, it needs to be solvable. And since solving the puzzle is how you progress, there’s less risk of solving the puzzle “too early”. Whenever you solve the puzzle, that is the climax.

You also don’t have to worry about rules like Knox’s Ten Commandments and variants thereof. The rules are codified into the game’s systems, as a matter of course. Each game can have a different set of rules, as appropriate to the particular story they’re trying to tell.

One of the best solvable mystery frameworks is the one established by The Return of the Obra Dinn, and further developed by The Case of the Golden Idol. These games feature lots of little mysteries, such as, how did each crew member of the Obra Dinn die? The answers are constrained to multiple choice lists. But the problem with multiple choice is that you can solve them with brute force, or at least rule out certain answers by just trying them. Obra Dinn addresses this problem by validating answers in groups of three. Once you have correctly solved three mysteries, then it will finally verify that you are correct. So if you reached the wrong conclusion about any of the mysteries, you get to try again, without having anything spoiled (quite unlike books). And if you’re truly stuck, some limited brute forcing is still available.

Of course, just like mystery novels, not all mystery games are trying to be “solvable”. Mystery games can be adventures, procedurals, or character studies. There is more to the mystery genre than the whodunnit.

The idea of solvability looms large above the mystery genre. Solvable books require setting expectations, and expectations tend to be elevated practically into moral law. But expectations can also be constraining and constricting. To get the most out of the mystery genre, we must understand how to condition our expectations on what each individual story is trying to do.


  1. Rob Grigjanis says

    My ideal mystery, is one that features a series of wacky suspects

    Brings to mind the film The Usual Suspects. At the end, you’re left with more puzzles than you started with, but a helluva ride.

  2. Tony Kehoe says

    The Ellery Queen novels were meant to be solvable, to the extent that they had an insert just before the solution was revealed stating that the reader was now in possession of all the clues necessary to solve the crime.

  3. says

    The concept of a mystery story being solvable makes me think of the Encyclopedia Brown books, because the whole gimmick there is that each chapter is a different mystery that Encyclopedia has to solve, and at the end of the chapter it stops and asks the reader what the answer is, and then you flip to the back of the book where it’s explained. I loved all those books when I was a kid, but the thing I think about now is that a lot of the “answers” didn’t make sense. Often it would be like, there’s some minor logical contradiction in one character’s alibi, and therefore they WERE LYING ABOUT THE ENTIRE THING.

    Some of the more ridiculous ones have stuck with me:

    There was one story where a man had big posters of playing cards in his house, and also he had 4 sons and 1 of them didn’t have a moustache, and THEREFORE the money was hidden behind the poster of the jack which doesn’t have a moustache. Clearly. Yeah. Don’t know how I missed that.

    There was a story where a woman was beat up in the bathroom of a restaurant, and everyone was like, well it couldn’t possibly have been a woman who did it, because the victim was really strong and surely no woman could beat her up, but it couldn’t have been a man, because that would have caused a commotion if a man was in the women’s bathroom. I remember thinking, wouldn’t it cause a commotion if someone is beating up someone else in the women’s bathroom, regardless of gender??? Anyway the answer was that it was a man dressed as a woman, and Encyclopedia’s assistant Sally figured this out because she saw a couple who weren’t sitting in their correct gender-role spots around the table at the restaurant (??? I don’t even know, these books were written in the 50s or something) and Encyclopedia and the other boys couldn’t figure it out because boys don’t know the rules of restaurant etiquette about who is supposed to sit where on a date (???).

    And there was the story which totally hinged on the fact that one character said he was going to take the elevator so he pushed the “up” button, and that meant that he WASN’T actually on basement level 2 where he thought he was- he was actually on basement level 1, which would have an “up” and “down” button. If he was really on basement level 2, he would have just said he pushed the button, not he pushed the up button. Yes seriously that was the answer.

    So overall the books were fun but you can’t take them that seriously because some of the answers are just really out there.

  4. Dunc says

    There’s a story from the making of The Big Sleep (Bogart / Bacall)… It’s quite a complicated Raymond Chandler story, in which a lot of people get killed for various complicated reasons. At one point, the production team realised that they didn’t understand one death in particular – a chauffeur. So they called Chandler up and asked him who killed the chauffeur, and why. His answer? He didn’t know either!

  5. Dr Sarah says

    Slight tangent: I remember that one of the speakers we had when I was in our university’s Medical Students’ Society was a private detective, and one of the questions someone asked in the Q&A was whether she could always figure out who dunnit in thrillers. She laughed and said that yes, she could, but it had nothing to do with her job; in thrillers it’s always the person you least expect, whereas in her job it normally ended up being exactly the person you thought it was likely to be from the start.

Leave a Reply

Your email address will not be published. Required fields are marked *