Novel Development Coach

Honest read: Your premise is strong: a coastline that erases itself is a literal image of a state erasing its atrocity, and giving that job to a mapmaker ties her profession directly to the theme. The risk is that the premise already tells the reader what happened (the crown did it), so if your Act 2 is built on discovering that, it has no real mystery, and that is almost certainly why it sags.

Caveat: I haven't read your pages, so the Act 2 diagnosis below is based on the most common causes for a premise like this. Check each one against your draft; don't assume.

1. Diagnosis

Your premise in one line: An exiled mapmaker must chart land that won't stay charted and finds that the disappearing coast is hiding the crown's crime.

Biggest strength: The central image does thematic work. Mapping means recording truth, and the coast undoes that recording every night. Most fantasy premises need a plot to reach their theme. Yours has the theme built into the setting.

Weakest link: the protagonist isn't implicated in the mystery. As pitched, Yeva is a witness to the crown's crime, not a participant in it. That causes three problems:

• The central question ("did the crown order a massacre?") is answered in the pitch, so Act 2 has no suspense engine.
• "Morally grey" is claimed but not built in. Being disgraced isn't the same as being grey.
• She has no personal cost in uncovering the truth, so she has no reason to hesitate, stall, or fight herself. Without that internal conflict, a middle turns into a list of clue-finding scenes.

The fix in one sentence: make the truth cost Yeva something personally. Ideally, she helped cause the massacre.

2. Why Your Act 2 Sags: Likely Culprits

Check these in order. They are ranked by how often they cause "slow and directionless" in a draft like yours.

1. The investigation loop. Chart the coast → it erases → find a fragment → repeat. Each cycle may be atmospheric, but if chapter 14 could swap places with chapter 17 without breaking anything, you have repetition without escalation.
2. No active goal after the setup. In Act 1 her goal is simple: do the job. If she spends Act 2 noticing things instead of pursuing something specific and measurable, the story drifts.
3. No pressure from antagonists. The crown is far away. Who is on the coast making her life harder this week? Your Goblin Emperor comp promises intrigue, but exile can remove every political opponent from the page.
4. The erasure has no rules, or its rules never change. If the vanishing works the same in chapter 20 as in chapter 5, the reader stops being unsettled. Piranesi's House keeps revealing new facts about itself. Your coast needs to do the same.
5. A soft midpoint. If your midpoint is "she finds more evidence," it adds information without changing direction. A real midpoint turn changes what the protagonist is trying to do.

Audit your draft in two hours

Make a scene spreadsheet for Act 2 (roughly 25k–70k words) with these columns:

• Any scene where "What's different after?" is blank or only says "she learns X": cut it, merge it, or make it cost her something.
• If most outcomes are a clean "yes," that's your sag. Middles run on "yes, but" and "no, and" outcomes: she gets what she wants, but it makes things worse, or she fails and things get worse anyway.
• Count your reversals. You want a real turn roughly every 8–12k words.

3. Strengthening the Midpoint: Three Options

A strong midpoint does one or more of these: reframes the mystery, changes the protagonist from reactive to proactive, or makes retreat impossible. Here are three options using your material, from lightest to most transformative.

Option A: The Complicity Reveal (my recommendation)

Yeva finds one surviving fragment of an old survey map, and it is in her own hand. Years ago, as the rising royal cartographer, she charted the coastal villages: their wells, their paths, which houses held how many people. She thought it was a census. It was a targeting map.

• The question changes from "What did the crown do?" to "What did I do, and can I bear to prove it?"
• This makes her morally grey in her choices, not just in her reputation.
• Act 2B gains a strong internal conflict. Every piece of evidence against the crown also condemns her. Now she has a reason to stall, hide things, or even help with the erasure, which gives you real tension for the next 25k words.

Option B: The Rule Reversal

She realizes the coast doesn't erase itself at random. It erases whatever she has charted. Her maps are the mechanism of the cover-up. The crown didn't exile her. It deployed her.

• This turns her greatest skill into the weapon used against the truth.
• Act 2B becomes a puzzle: how do you record a truth without mapping it? Through testimony, songs, bodies, or salt? The magic system now carries the theme.
• It combines well with Option A.

Option C: The Offer

An envoy arrives from court with a pardon. Her title and home will be restored if she certifies that the coast is uninhabited and always was.

• This gives the political thread a face and a deadline, bringing back the intrigue your comps promise.
• It works as a false victory: everything she wanted in Act 1, offered at the worst possible moment.
• It is the weakest option on its own, but it's a strong Act 2B complication if you also use Option A.

My pick: Option A at the midpoint, then Option C about 10k words later as the pressure that squeezes her. If Option B appeals to you, bring it in as the Act 2 low point, where she realizes she has been erasing the evidence all along.

4. Sample Act 2 Beat Sheet (targeting 104k words)

Note on length: you're at 92k and aiming for 100–110k. Don't add words to Act 2. It will probably shrink by 5–10k once you cut the loop scenes. The new words should go into rebuilding the midpoint, adding the warden and envoy thread, and giving Act 3 room for consequences.

5. Scene Example: The Midpoint (Option A)

Sample prose to show how the turn might land, in close third past:

The fragment had survived because someone had used it to plug a crack in a cistern wall, folded eight times and packed with pitch. Yeva worked it open with the flat of her knife.

A village. Twelve houses, a well, the goat path down to the shingle. Each roof marked with a small numeral—4, 6, 3, 7—in the careful serifed hand she'd been praised for at nineteen. She had thought them a tithe count. She remembered the commission: an inventory of His Majesty's coastal subjects, for their better protection. She remembered being proud of the sevens, how she'd learned to close the stroke so it could not be mistaken for a one.

Behind her, Dessa asked what she'd found.

"Nothing," Yeva said, and her hands were already folding it smaller.

Why it works:

• The revelation comes through her expertise. She recognizes her own sevens, and only she could.
• It's grounded in a detail (pride in a pen stroke) rather than stated emotion.
• The last line is a choice, not a reaction. She lies right away, and that lie becomes the engine of Act 2B.

Next Steps

1. Today (2 hours): run the scene audit. Build the spreadsheet for everything between your Act 1 break and your current climax. Highlight every scene where nothing changes. Don't cut yet; just count them. That number is your sag, measured.
2. This week: choose and draft the midpoint. Pick Option A, B, C, or your own version. Write that one scene cleanly (1,500–2,500 words) before touching anything else. Every Act 2B scene will be revised to follow from it, so it needs to exist first.
3. Next: outline Act 2B again from scratch. Use the beat sheet as scaffolding, and give every scene card a goal and a "no, and" or "yes, but" outcome. Plan to cut or merge the loop scenes, and give the warden and envoy at least four on-page confrontations between them.

Follow-up question: In your current draft, why was Yeva disgraced, and does it have any connection to the massacre? If there's no connection yet, that's the most valuable thing we could change. If there is one and you're saving it for the ending, it may be the midpoint you already have and are holding back.