What I Look For in the First 50 Pages of a Fantasy Manuscript
Short answer
In a fantasy sample I read for eight things, roughly in this order: whether the opening promise matches the book, how many invented proper nouns arrive before the reader cares, whether the magic is demonstrated or explained, whether the prologue earns its place, whether chapter one starts in the right chapter, whether something irreversible happens by page fifty, the sentence-level habits, and whether a polished opening is hiding an unrevised middle.
By page fifty of a fantasy manuscript I can usually tell you how the rest of it goes, which sounds like I’m bragging but honestly, it’s more of an occupational hazard than anything else. The opening pages are where a writer makes every decision that the remaining four hundred will be spent clarifying, or walking back, and those decisions are visible from the outside in a way they rarely are from the inside.
Authors ask me what I’m looking at when I read a sample, usually because they’re hoping the answer is prose, and the prose is the part they have polished. The prose matters, of course, but it’s maybe fifth on the list. What follows is closer to the real order.
Does the opening promise match the book?
Fantasy runs on an agreement made early and honored late. A reader who opens to a rain-soaked funeral and a boy being handed his father’s sword has been promised a particular kind of book, and if page two hundred delivers a court-intrigue novel about grain tariffs, the reader is absolutely not delighted by the surprise, they’re annoyed that they were reading the wrong book for six hours.
So the first thing I do is write down, in one sentence, what the opening seems to be promising. Then I check that sentence against the synopsis. When those two things disagree, almost everything else in my report follows from the disagreement, and no amount of line-level work is going to fix it.
How many invented proper nouns arrive in the first ten pages
This is the most common failure I see in epic fantasy and the easiest to diagnose, because you can count it. I mark every invented proper noun in the first ten pages: names, places, houses, gods, orders, magic systems, wars, currencies. When that number climbs past about a dozen, a reader who isn’t already invested has nothing to attach any of them to, and the names become meaningless.
The fix is almost never cutting the worldbuilding, which is usually good and is usually why the book exists. The fix is delaying the introductions until the reader has a reason to care. A name given to a reader who’s invested is vital information; the same name given to a reader who doesn’t yet care yet is just annoying homework.
Since it’s easier to see than to describe, here is the same opening beat written twice. Both are invented, so nobody’s draft is being held up as a cautionary tale.
Ten names:
Kaeryn came down out of the Sablemark with the last of her father’s coin sewn into her hem, forty threnn against a winter that had already killed better-provisioned people than her. Ashgate sat below in its ring of grey smoke, and past it the towers of Vosh Anduel, where the Hollow Accord had been signed and broken inside the same year. Her mother had been Vaelric before the marriage, which meant nothing now that House Vaelric was struck from the rolls, though it had meant enough during the Ninefold War that a woman with those cheekbones still drew looks in the lower market. She touched the pendant of Tessarine at her throat, a habit she had never managed to break, and started down.
Two names:
Kaeryn came down out of the mountains with the last of her father’s coin sewn into her hem, forty pieces against a winter that had already killed better-provisioned people than her. Ashgate sat below in its ring of grey smoke. Her mother’s family had been struck from the rolls the year Kaeryn was born, which meant nothing to anyone now, though it had meant enough once that a woman with those cheekbones still drew looks in the lower market. She touched the pendant at her throat, a habit she had never managed to break, and started down.
The second version is eleven words shorter and carries the same information. A woman is walking into a city she can’t afford, from a family that used to matter, wearing something she’s attached to but doesn’t yet explain. Everything a reader needs in order to lean forward is still on there.
What’s gone now is what we need to talk about, because nothing was actually cut. The mountains keep their name the first time Kaeryn has a reason to say it aloud. The currency becomes threnn at the moment somebody refuses to take it. The Hollow Accord arrives when a character’s safety depends on it having been broken. The goddess gets her name the first time Kaeryn asks her for something and isn’t answered, which makes it a scene rather than a label. The war can wait until a veteran of it is standing in the room.
Each of those names should be doing something at the moment it arrives, instead of asking the reader to hold it in escrow against a payoff two hundred pages out. And the worldbuilding doesn’t suffer. It’s the same world; the reader is simply being handed the information in a way they can remember.
The count is a diagnostic, not the rule. If your first ten pages have nine invented names and every one of them is introduced in context, keep them. If there are nine and four could be swapped for “the mountains” or “the city” without a reader losing anything, those four are just decoration.
Does the magic system demonstrate itself, or explain itself?
I look for the first moment the magic system appears and then I look at what form that appearance takes. If a character explains the rules to another character who already knows them, the manuscript has an exposition problem that will recur, and I will find it again around the midpoint, and again before the climax.
The version that works is a rule the reader learns by watching it cost somebody something. A magic system that hasn’t upset somebody is not yet a magic system, it’s just a list of rules.
Speaking of rules, I’m also checking whether the rules stay put. Internal consistency is the thing readers of this genre notice most reliably and forgive least, and inconsistency that starts in the first fifty pages is usually structural rather than careless, which means it’s worth catching before a copyedit rather than during one.
Does the prologue earn its place?
I am not against prologues, and I’ll defend a good one against anyone, including the many agents who have declared them dead. But I ask the same question every time, which is whether the prologue is doing work that chapter one could do, or whether it exists because the author wasn’t sure the opening was interesting enough on its own.
A prologue that dramatizes an event the protagonist doesn’t know about, and that pays off later, works. A prologue that summarizes ten thousand years of history in the voice of no one in particular is a clear sign that the first chapter starts in the wrong place.
Whether chapter one starts in the right chapter
Which brings me to the thing I find in a surprising number of first novels: the real beginning is somewhere in chapter three. Not always, but often enough that I look for it deliberately.
The tell is usually a shift in the writing itself. Somewhere the sentences get faster, the character stops being described and starts being someone, and the author clearly began enjoying themselves. That point is frequently where the book should open, with the preceding material either cut or redistributed into the places it becomes relevant.
Authors hate this note. They hate it because those first chapters were legit hard to write, and because writing them was how the author found the character. Both of those things can be true and the chapters can still be scaffolding.
What has to have happened by page fifty
Something irreversible. Not necessarily huge, and certainly not necessarily violent, but something that means the protagonist absolutely cannot go back to the life they had on page one.
I flag manuscripts where page fifty could be deleted along with everything before it and chapter six would still make sense. That’s not a pacing problem exactly, it usually just means the author was building confidence before starting the actual book.
The sentence-level habits I read for
Now we get to prose, and even here I’m reading for patterns rather than individual sins, because a sample tells me what the manuscript does habitually.
I watch for filter words, the she saw and he felt and she noticed that insert a pane of glass between the reader and the moment. I watch for adjectives arriving in pairs, which is almost always a sign the writer didn’t trust the first one enough to do the job. I watch dialogue tags, because a manuscript that says he exclaimed on page four will go there three hundred more times. And I watch for the rhythm of the paragraphs, since a writer who has only one sentence length available is going to struggle to control tension later, when they need the pace to change.
None of these are fatal and all of them are fixable, which is exactly why I want to see them early. A copyedit that finds a habit on page thirty can build the style sheet around it. A copyedit that discovers it on page three hundred has been inconsistent for two hundred and seventy pages.
When the first fifty pages are good and the book still isn’t
This happens, and it’s the case authors are least prepared for. An opening can be polished to a shine because it’s been revised twenty times, workshopped, read aloud at three events, and rewritten after every piece of feedback, while chapters nine through twenty have been read by exactly one person.
So when a sample is unusually clean, I ask what the middle looks like, and I ask how many times the opening has been revised. If the answer is a number the author is proud of, we probably need to talk about the middle before we talk about anything else.
What this means if you’re about to hire an editor
If you’re choosing between an editorial assessment and jumping straight to a copyedit, the questions above are a decent self-diagnostic. Where the promise matches the book, the magic costs something, the opening starts in the right place, and something irreversible happens early, you’re likely ready for sentence-level work.
Where two or more of those are shaky, a copyedit will produce a beautifully punctuated version of a structural problem, and you will have paid for the wrong pass. I’d much rather tell you that before you pay me than after.
I read the first few pages of every manuscript I’m asked to quote on, and I’ll tell you which pass I think you need even when it’s the cheaper one.
Common questions about fantasy manuscript openings
How many pages does an editor actually need to assess a manuscript?
Fifty is the number I use, because it’s long enough to show habits rather than one good day of writing, and short enough that patterns are still patterns rather than plot. Ten pages tells me about the prose. Fifty tells me about the book.
How many invented names is too many in a fantasy opening?
Past about a dozen in the first ten pages, a reader who isn’t already invested has nothing to attach them to. The fix is rarely cutting worldbuilding, it’s delaying the introductions until the reader has a reason to care which house somebody’s mother married out of.
Should my fantasy novel have a prologue?
It should if the prologue dramatizes an event the protagonist doesn’t know about and pays off later. It shouldn’t if it summarizes history in the voice of nobody in particular, which usually means chapter one starts in the wrong place and the prologue is covering for it.
What are filter words and why do editors flag them?
Filter words are the she saw, he felt, she noticed constructions that put a narrator between the reader and the moment. They aren’t errors, and one here and there is fine. A manuscript that uses them habitually is holding the reader at arm’s length for four hundred pages without meaning to.
Do I need an editorial assessment or can I go straight to a copyedit?
If the promise matches the book, the magic costs something, the opening starts in the right place, and something irreversible happens early, you’re likely ready for sentence-level work. If two or more of those are shaky, a copyedit will give you a beautifully punctuated structural problem. Rates for both are on my services page.