The difference between a twist and a cheat

A twist and a cheat feel identical for about three seconds. Then the reader thinks back over what just happened, and one of two things happens. Either the pieces click into a shape that was there the whole time, or the reader realizes you never gave them the pieces at all. That gap, between recognition and betrayal, is the entire difference. It has nothing to do with how surprising the reveal is.

What a fair twist actually requires

A twist is fair when a careful reader, paying attention on a first read, could have arrived at the answer before the narrator announces it. Not a reader who rereads the book three times hunting for clues. Not a reader who happens to share the author's specialist knowledge. An ordinary attentive reader, moving through the story once, at story speed. If that reader could not have gotten there with the information you gave them, you haven't written a twist. You've written an ambush.

This is the standard detective fiction settled on a century ago, and it still holds for every genre that uses a reveal. The clue has to be visible. It doesn't have to be obvious: in fact it works better disguised as something else, a detail that reads as scene-setting until it snaps into evidence. But it has to be on the page, in a scene the reader actually witnessed, not summarized after the fact as something the detective “had suspected for some time.”

What makes a cheat

A cheat withholds information the point-of-view character clearly possessed. If your narrator walked past the murder weapon and thought nothing of it, fine, that's a plant. If your narrator saw the murder happen and you simply didn't tell the reader because the sentence was inconvenient to your plot, that's not suspense. That's lying to the reader in a way only the author benefits from. The other common cheat is the invented solution: something that shows up in the last chapter with no earlier appearance, existing only to solve a problem the plot painted itself into. Deus ex machina isn't just a term for gods on wires. It's any resolution that required nothing the story had already established.

The test to run before you commit to a twist

Before you write the reveal scene, go back through the manuscript and list every fact the reveal depends on. For each one, find the earlier scene where it appears. If you can't find one, you have two options: write it in now, or admit the twist doesn't work yet and needs a different shape. Stated flatly, this sounds obvious. It is astonishing how often writers skip it, because the reveal felt so satisfying in their head that they never checked whether it existed on the page.

A sharper version of the same test: read the chapter containing the reveal as if you were a stranger, then ask whether the clue, on its own, would have meant anything to that stranger the first time through. A clue that only makes sense in hindsight isn't a clue. It's an alibi you're writing for yourself after the fact.

Plant (fair)
  Ch. 3: the housekeeper mentions she throws out the mail before
  the family sees the post, so Mr. Voss doesn't fret over bills.
  The reader files this as characterization.

Payoff (earned)
  Ch. 19: the letter that would have warned Voss never reached
  him. The housekeeper's habit, established sixteen chapters
  earlier, is the reason. Nothing new is invented.

Cheat (unearned)
  Ch. 19: Voss reveals he always suspected the letter was
  intercepted, a suspicion never shown, never hinted, never once
  present in his narration until the sentence that needs it.

The POV problem is the usual failure point

Most cheats aren't invented maliciously. They come from close POV without discipline. Deep third or first person locks the reader to a single consciousness, and that consciousness knows things. The moment your POV character notices, remembers, or reacts to information the narration then hides from the reader, you've broken the contract close POV runs on. The character's thoughts are the reader's only window into the scene. If those thoughts skip past the truth on purpose, the fix isn't clever. It's a trick, and trained readers spot the seam even when they can't name it.

Before you write the reveal, it's worth checking whether the detail actually made it onto the page somewhere the reader could have caught it, not just in the outline of what you meant to include. Chaptered's foreshadowing tracker exists for exactly that gap: log a planted element against the scene where it's dropped, mark whether it's still planted, resolved, or abandoned, and tie it to the scene where it eventually pays off, so the check happens against the record instead of your memory.