"Sometimes it's not that deep," aka, How to Write a Mystery (That's Actually There)
A good rule of thumb about predicting a well-written mystery is that other people are predicting the same things. Even better if a lot of other people, on their own, are analyzing the material closely and each, on their own, are reaching the same theories and conclusions! Yes, this can mean everyone's falling for the same red herring too, of course, but generally speaking, no writer is putting all that effort into writing a mystery only one person can solve.
The thing is, a good mystery should be predictable, but at varying speeds based on the attentiveness of the audience.
The most laser-focused readers (or viewers), the true Holmesian mystery-solving mystery readers with encyclopedic knowledge and a fine-toothed comb to go through the material, should be ahead of the characters trying to solve the mystery, but not so far ahead that the story becomes boring, ideally. Your mystery shouldn't be so predictable that it can be figured out on page one but, if a truly clever sleuth does figure out your mystery on page one, your story should also be good enough that it is still an enjoyable tale even if the audience know how it ends.
The attentive reader should figure out the mystery at about the same time as the main character. They should have the "Aha!" moment about two seconds before the character solves it for that little thrill of feeling very clever, which to authors is very funny, because if we're doing our job right, we didn't just set the clues for you to solve the mystery, we agonized over the timing of each clue's introduction just to make sure you got that little thrill of self-satisfaction at the moment we wanted you to.
The casual reader should figure out the mystery when the characters figure it out. They should get the enjoyment of seeing the mystery solved, and in retrospect be able to see the clues, but in the meantime have enjoyed their time nonetheless. It should be a good story on its own.
Now, this is relevant to fandom theory mongering because clues are not accidental. And there are usually a lot of them, if the author is doing their job. Yes, you should be able to analyze the costume colors to see that there are parallels between the green dress or whatever that the lead is wearing in this episode and how it's a call back to another episode and that means blahblahblah is going to happen, but, that sort of deep-cut, freeze-frame sleuthing is usually only in support of more overt clues intended for general audiences.
(Not to pick on any one fandom, but Ted Lasso S3 for example had a lot of Tedbecca shippers looking for hidden clues in the cinematography as their hopes faded for a canon confirmation of their ship. Alas, those didn't bear out, because they were not accompanied by textual evidence in addition to the subtextual evidence of how any moment now, these characters are going to stop dating and pursuing completely different people and actually hook up with or even verbally express confirmed interest in each other.)
Of course, a mystery can be tough to solve and be satisfying! It is sometimes even possible to solve those truly out there or even not yet fully supported mysteries that eventually turn out to be true (say, predicting a later book in the series before all the clues are even there from installments in the meantime). It can be really satisfying to correctly extrapolate from incomplete data before the author even intended you to see it or had figured it out themselves!
However, more often than not, the clues are deliberate and to assume you're the only one seeing them is probably a sign that they're not actually there.
A good mystery, a well-written one, should provide all the clues for the audience to solve the mystery on their own, within the text, even if the last few pages are ripped out. A good mystery is not a "gotcha". And a sign of a good mystery is that more than one person can pick up on all the clues because those clues are placed deliberately and yes, that requires a certain amount of sign-posting as well and in fact, one of the most fiendishly difficult things to do as a writer is strike the proper balance between sign posting your clues at just the right time to reward your attentive audience with enough data to solve the mystery just before the characters do, but not so early that they lose interest.
So, why is this important for fandoms? Because it's entirely possible to go down the conspiracy theory rabbit hole and convince yourself of stuff that's not there and then get really bummed when it doesn't happen. We're all guilty of this, myself very much included, especially in the largely subtextual world of slash shipping.
There's also real world exceptions and extenuating circumstances to my statement that a good mystery will actually provide the clues: like studio changes, dropped threads due to contractual complications, or a crowded writers room where not every idea gets pursued.
Plus, we've got the whole goddamn JJ Abrams-inspired mystery box bullshit run of television making creators think it's cool to pull one over on the audience and deny the clues they set up just to make a "twist" that no one predicted because it wasn't there, thus betraying the principles of a good mystery, and that also muddies the waters about how good mysteries should be written. (If you can't tell, I absolutely despise writers who pull this bullshit with the intent to trick their audience and they do not deserve any praise for being "clever" to just surprise people with an outcome that's not supported by previous text ala "rocks fall, everyone dies", ugh.)
But besides the damned mystery box crud, there are some good rules of thumb for determining if the clues you're picking up are real or not:
Is the subtext supported by text? Not "is her green dress backed up by a micro-expression glance you can barely see if you freeze-frame the show" but actually in the text. Does someone say, in dialogue, "Hey, we should go on a date," when you're reading subtext between two characters and can you be certain that the writers intended that text to be read the way you read it? One way to tell is if this sort of thing happens more than once, if there are more clues. Of course, there's also red herrings, etc, and hey, that's half the fun of a mystery, not all clues are Clues. But generally speaking, there's more than one clue for important stuff.
Is there evidence against your theory and have you considered it? This is basic Logical Thinking 101, of course, but only looking for evidence that confirms your theory and ignoring everything that doesn't will definitely convince you of some pretty wild stuff in a hurry!
Have you considered the genre of the work and whether it even has mysteries? Or is it, for example, a comedy which might address those "clues", most likely comedically, but might just as easily not even realize they were seen as clues?
But mostly importantly: do other people have this theory? Did they arrive to it on their own, based on the same text? If you are the only person with a theory and cannot use textual evidence to convince more than one person who is not of the same background as you (ie, convince your less-progressive dad that these two guys definitely want to boink based on the way they look at each other) there is a very likely chance you are stringing together disparate data points to match your desires, not the evidence. Bonus points if, seriously, you can convince someone who is of the same demographic as the content creators (writers, directors, etc.) that these clues are intentional. Most likely, your less-than-progressive dad has more in common with the thought process that went into creating the vast majority of content than your average Tumblr user does. (OFMD is the exception that proves the rule.)
As a final note, one thing it's very dangerous to do as a writer is introduce the idea of clues and a mystery without making your mystery air tight. Mysteries invite the reader to put their brain on high alert. It means they're looking for clues everywhere and even seemingly innocuous throw-away descriptions or the camera lingering on a random prop for too long can be misinterpreted and end up pissing off the audience who thinks it's a Clue.
You have to be so deliberate with a mystery, because it's not about just controlling the clues you release, but controlling the perception of the audience so they know what isn't a clue (unless it's a red herring, of course). That's part of why scifi author David Brin said all aspiring writers should have their first novel be a mystery, because it teaches you volumes on how to control your narrative and make everything intentional.
