Why Endings Feel Unearned: The Setup-Payoff Contract
There is a difference between a reader who closes your book and says "I didn't see that coming" and one who closes it and says "that came out of nowhere." The first response is a compliment. The second is an autopsy. Both describe surprise, but only one describes a story that honored its contract with the reader.
That contract is not a formal document. It is assembled invisibly across your first fifty pages through every choice you make about whose pain we witness, what dangers are introduced, which questions are asked without being answered, and what kind of story the prose itself seems to be promising. A thriller that opens with slow domestic dread is promising something different from one that opens with a body on the first page. A literary novel that introduces a protagonist haunted by her mother's silence is promising, implicitly, that the silence will break—or that breaking it will cost something irreversible.
The problem is that writers make these promises in a state of creative excitement, often without tracking them. By the time you reach your climax, you may have forgotten what page twelve was guaranteeing. Your ending may be emotionally satisfying to you because you know everything that happened in the drafting process. Your reader only knows what made it onto the page—and what your opening pages told them to expect.
This is where AI becomes a genuinely useful revision partner. It has no access to your intentions. It only has your text. That limitation, which can frustrate writers in other contexts, is exactly the right quality for auditing whether your ending was earned by your own setup.
The Promise Inventory: Extracting What Your First 50 Pages Guarantee
Before you can audit your climax, you need a complete record of what your opening has already promised. This is harder than it sounds. Promises exist at multiple levels simultaneously: plot-level stakes (what could be lost or gained), emotional stakes (what the character is afraid of or desperately needs), thematic questions (the abstract argument the novel is making about how the world works), and tonal promises (the emotional register and genre expectations the prose style itself creates).
Most writers can articulate the plot promises easily. The thematic and emotional ones tend to live in the subtext, which is precisely why they're easy to abandon without noticing.
The following prompt gives AI the structural task of surfacing all four layers from your opening pages:
I'm going to paste the first 50 pages of my novel manuscript. Your task is to conduct a Promise Inventory—a systematic extraction of every contract these pages make with the reader, whether the writer intended it or not. Organize your analysis into four categories: 1. PLOT PROMISES: What specific outcomes, confrontations, or revelations do these pages make the reader expect? Include both explicit setups (a named antagonist, a stated goal) and implicit ones (a situation whose natural trajectory points toward a specific type of conflict). 2. EMOTIONAL PROMISES: What does the protagonist's wound, fear, or core desire guarantee the reader about the kind of emotional experience this novel will deliver? What internal transformation is being foreshadowed? 3. THEMATIC PROMISES: What questions about human experience are these pages raising without yet answering? State each as a question the novel appears to be building toward answering. Include thematic contradictions or tensions that are introduced. 4. TONAL AND GENRE PROMISES: Based purely on what appears in these pages, what kind of story does the prose style, pacing, and imagery suggest this will be? What expectations does this create about how the story will end—in terms of register, resolution type (ambiguous vs. definitive), and emotional destination? For each item, note the specific page, paragraph, or image that generates the promise. Do not speculate about the rest of the novel. Work only from what I give you. [PASTE YOUR FIRST 50 PAGES HERE]
Run this before you paste your ending. You want the inventory generated without the AI knowing where you ended up—otherwise it will unconsciously rationalize backward from the ending you gave it. Keep the output. This document is the standard against which your climax will be measured.
The Climax Audit: Mapping Your Ending Against the Inventory
Once you have your Promise Inventory, the audit itself is a gap analysis. You are asking: which of these promises are cashed out in the ending? Which are left uncollected? Which are collected in a way that technically closes the loop but emotionally cheats the reader because the resolution arrives from outside the logic of the setup?
The third category is the trickiest. A writer who realizes midway through a draft that their original thematic question isn't working may quietly shift to a different question—one that the ending answers beautifully. The ending is not incoherent. But it's answering a question the reader was never asked to hold. The result is an ending that feels somehow beside the point even when it's technically well-executed.
I'm going to give you two documents. The first is a Promise Inventory generated from my novel's opening 50 pages. The second is my climax and final chapter (or a detailed summary of each if the full text is too long). Your task is to conduct a Climax Audit with three outputs: 1. CASHED CHECKS: For each item in the Promise Inventory, identify whether the climax and ending fulfill it, and specifically how. Be precise—"the wound is addressed" is not sufficient. Describe the mechanism by which it's resolved. 2. ORPHANED SETUPS: Identify every item in the Promise Inventory that the ending does not address or resolve. Flag these as potential structural liabilities. Note whether the omission is likely to read as intentional ambiguity or as a forgotten thread. 3. IMPORTED RESOLUTIONS: Identify any element of the climax or ending—plot developments, character decisions, thematic conclusions—that does not appear to grow from material in the opening inventory. These are endings that may be internally satisfying but feel narratively unearned because the reader was never prepared for them. For each, describe what setup in the opening pages would have been needed to make this resolution feel inevitable rather than arbitrary. Do not tell me whether the book is good. Focus entirely on structural alignment between setup and payoff. [PASTE PROMISE INVENTORY HERE] [PASTE CLIMAX AND FINAL CHAPTER HERE]
The output from this prompt will often be uncomfortable. That discomfort is information. An orphaned setup is not a failure—it is a navigation point. You now know exactly where to look when readers tell you something felt missing without being able to name what.
The Inevitability Test: Stress-Testing Your Logic from Both Sides
One of the more useful analytical moves in fiction criticism is the dual-argument test: if you can construct a strong case for a position and you cannot construct a coherent case against it, you probably have something solid. If both arguments are strong, you have a genuine tension worth examining. If only the counter-argument holds up, you have a problem.
Applied to your climax, this means asking AI to argue two contradictory positions with equal seriousness: that your ending was inevitable given your setup, and that your ending came from nowhere. The goal is not to let AI render a verdict. The goal is to surface the specific evidence for each claim so you can evaluate it yourself.
I need you to stress-test the structural logic of my novel's ending using a dual-argument format. You will argue two positions in sequence, each with equal rigor and specificity. Do not favor one position or hedge toward a compromise. Make the strongest possible case for each. POSITION A — THE EARNED ENDING ARGUMENT: Argue that this ending was inevitable given the setup in the opening pages. Identify every piece of foreshadowing, character behavior, thematic development, and planted detail that, in retrospect, points toward exactly this conclusion. Describe the causal chain from opening promise to final payoff. Make a case that a reader who paid close attention could have anticipated this ending's emotional and thematic destination, even if not its specific events. POSITION B — THE UNEARNED ENDING ARGUMENT: Argue that this ending arrives from outside the logic of the setup. Identify every place where the ending requires something the reader was never given: a character capability not established, a thematic resolution not prepared for, an emotional shift that lacks sufficient groundwork. Describe specifically what a reader relying only on the text would have reasonably expected instead, and how the actual ending diverges from those expectations. After both arguments, list the three specific moments in the opening 50 pages that most directly determine whether this ending reads as earned or arbitrary. These are your leverage points for revision. [PASTE OPENING 50 PAGES + CLIMAX SUMMARY HERE]
What you are looking for in the output is not which argument "wins." You are looking for the specific textual evidence that Position B uses. If it cites things you cannot counter, those are your revision targets. If Position A's argument is detailed, specific, and traces a clear causal chain through the manuscript, you are likely in better shape than your anxiety about the ending suggests.
Revision Triage: Retrofitting Foreshadowing and Closing the Loop
Once you know what is missing, the repair work falls into three categories: adding foreshadowing to the early chapters, adjusting the midpoint to raise the stakes that your climax requires, or rewriting the denouement so that it closes the thematic question your opening pages asked. Each requires a different type of intervention, and AI can help you identify not just what needs to change but where in the manuscript the change belongs.
No comments yet. Be the first to comment!