{ "entity": "gemini-3-1-pro", "entity_type": "model", "category": "Book writing", "suite": "suite-2026-08-tasks4/book-writing", "model": "google/gemini-3.1-pro-preview", "judges": [ "anthropic/claude-sonnet-5", "openai/gpt-5.6-terra", "x-ai/grok-4.5" ], "judge_protocol": "judge-2026-08b", "anchors": "solo-2026-08-a", "score": 87, "caps": 0, "cost_usd": 0.3843, "started_at": "2026-08-16T18:15:07.833Z", "finished_at": "2026-08-16T18:16:19.704Z", "tasks": [ { "id": "bw1", "name": "Is it a book?", "prompt": "'Forty years driving a Blackpool taxi — everyone says I should write a book.' Everyone is sometimes right. Explain in plain terms what separates a book from a pile of anecdotes — the organising spine — then propose 3 different spines this one could hang on: one chronological, one thematic, one built around a single night. For each, a sentence on what it makes possible and a sentence on what it costs.", "output": "**The Organising Spine: What Separates a Book from a Pile of Anecdotes**\n\nA pile of anecdotes is like a handful of loose pearls: each one might be shiny and interesting, but if you hand them to someone, they will eventually drop them and walk away. A book is the string that runs through them, turning them into a necklace. \n\nIn publishing, that string is called the \"spine.\" It is the central argument, the emotional journey, or the structural framework that gives the reader a reason to turn the page. Anecdotes just say, *“Here is a crazy thing that happened.”* A spine says, *“Here is how this crazy thing connects to the bigger picture of who I am, what this town is, or how human beings behave.”* Without a spine, a reader gets exhausted after twenty pages of random funny stories because there is no sense of destination.\n\nHere are three different spines upon which forty years of Blackpool taxi stories could hang:\n\n### 1. The Chronological Spine: *The Rise, Fall, and Survival of the Seaside*\n**The Concept:** The book tracks a linear timeline from the driver’s first terrified shift in the booming 1980s to their confident, perhaps cynical, present day, mirroring the economic and cultural changes of Blackpool itself. \n* **What it makes possible:** This structure allows the reader to experience a grand, sweeping history of how both a famous British town and the narrator evolved over four decades.\n* **What it costs:** It forces you to bridge the gaps between the exciting moments, risking a sagging pace as you are obligated to explain the boring, transitional years just to keep the timeline intact.\n\n### 2. The Thematic Spine: *The Confessional in the Rearview Mirror*\n**The Concept:** Time is irrelevant; instead, chapters are divided by the types of human behavior witnessed in the back seat (e.g., The Runaways, The Drunks, The Broken-Hearted, The Dreamers). \n* **What it makes possible:** This approach allows you to cherry-pick only the absolute best, most emotionally powerful stories and group them together for maximum impact.\n* **What it costs:** It sacrifices the narrator's personal arc, often turning the driver into a static, unchanging observer rather than a living character who ages and learns.\n\n### 3. The Single-Night Spine: *The Last Shift*\n**The Concept:** The entire book takes place over one 12-hour shift—perhaps the driver’s final night before retirement, or the chaotic night of the Illuminations switch-on—with the events of the current night triggering flashbacks to the past forty years.\n* **What it makes possible:** This framing device creates an intense, ticking-clock momentum and a highly cinematic atmosphere that keeps the reader hooked from dusk till dawn.\n* **What it costs:** It requires heavily contriving the events of that specific night so that every modern passenger perfectly triggers a relevant flashback to the past.", "latency_ms": 17778, "cost": 0.022796, "task_score": 9, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 8, "violates": false, "reasoning": "Clear, accurate explanation of spine concept; three distinct spines each with a benefit/cost sentence as requested. Well-organized, concise, useful. Minor stylistic embellishment but meets all instructions." }, { "judge": "openai/gpt-5.6-terra", "score": 10, "violates": false, "reasoning": "Clearly explains the organising spine in accessible language and provides exactly three distinct, requested structures. Each includes a concept plus clear sentences on what it enables and what it sacrifices; all are well suited to the premi" }, { "judge": "x-ai/grok-4.5", "score": 9, "violates": false, "reasoning": "Clear, accurate explanation of spine; three well-differentiated proposals each with exact required sentences on gains and costs. Minor polish only." } ], "machine_checks": null, "machine_violation": false, "machine_reasons": [] }, { "id": "bw2", "name": "Twelve chapters, one year", "prompt": "Outline 'A Year on the Plot' — a beginner-friendly book about a first year on an allotment — as exactly 12 chapters, numbered 1-12, one line each and no sub-points: chapter title plus what the reader can DO after reading it. Keep a practical thread throughout, and the January chapter must not pretend January is exciting.", "output": "1. January: The Cold, Muddy Reality — You will be able to map your plot dimensions on paper, order seeds from the warmth of your sofa, and resist the urge to trudge through freezing mud when there is absolutely nothing to plant yet.\n2. February: Clearing the Decks — You will be able to safely clear winter debris, repair broken raised beds, and chit your first seed potatoes on a bright indoor windowsill.\n3. March: Waking Up the Soil — You will be able to weed out early invaders, dig in organic compost, and direct-sow your first hardy crops like broad beans and radishes.\n4. April: The Spring Sowing Sprint — You will be able to set up a simple crop rotation system, sow the bulk of your root vegetables, and protect tender seedlings from late frosts.\n5. May: Planting Out and Pacing Yourself — You will be able to harden off indoor-grown plants, safely transplant them into their final outdoor positions, and establish a sustainable weekly watering routine.\n6. June: Battling Bugs and First Bites — You will be able to identify and organically manage common allotment pests, keep on top of rampant weeds, and harvest your first early salads.\n7. July: The Watering and Weeding Treadmill — You will be able to implement water-saving mulching techniques, pinch out tomato side-shoots, and harvest courgettes before they turn into giant marrows.\n8. August: The Heavy Harvest — You will be able to correctly lift and cure onions, freeze or store your gluts of summer produce, and sow quick-growing autumn catch-crops.\n9. September: The Autumn Shift — You will be able to clear spent summer crops, plant overwintering garlic sets, and build a functioning compost bin using your seasonal green waste.\n10. October: Putting the Plot to Bed — You will be able to harvest your maincrop potatoes, sow green manures to protect bare soil, and cover unused beds with cardboard to suppress winter weeds.\n11. November: Trees, Bushes, and Boundaries — You will be able to plant bare-root fruit bushes, prune dormant apple trees, and reinforce your sheds or netting against winter storms.\n12. December: Tools, Tidy-Ups, and Taking Stock — You will be able to clean, sharpen, and oil your hand tools, review your first-year crop successes, and leave the plot alone to rest.", "latency_ms": 19144, "cost": 0.029536, "task_score": 8, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 8, "violates": false, "reasoning": "Meets format, 12 chapters, practical actions, January appropriately unglamorous. Titles include colons which slightly exceed 'title plus what reader can do' simplicity but still one line, no sub-points." }, { "judge": "openai/gpt-5.6-terra", "score": 8, "violates": false, "reasoning": "Meets the exact 12-line, numbered, no-subpoint format and keeps January realistically subdued. Practical actions are clear and seasonal. Minor factual overstatement: January is not universally a month with “absolutely nothing” to plant, dep" }, { "judge": "x-ai/grok-4.5", "score": 10, "violates": false, "reasoning": "Exact 12 one-line chapters with titles and actionable outcomes; practical focus; January realistically dull as required." } ], "machine_checks": { "exact_bullets": 12 }, "machine_violation": false, "machine_reasons": [] }, { "id": "bw3", "name": "Back cover, no spoilers", "prompt": "From this summary, write back-cover blurb of max 100 words: 'Retired lighthouse keeper Edith Brand receives letters addressed to her dead sister, postmarked from the village where they grew up — a village drowned by a reservoir in 1971. She returns, rents a room above the pub from a landlord who remembers too much, and discovers her sister's death forty years ago may have been neither accident nor suicide. The truth involves their father.' The blurb must NOT reveal the father's involvement — sell the mystery, don't solve it. Return only the blurb.", "output": "Retired lighthouse keeper Edith Brand thought the past was buried. Then the letters arrive. \n\nAddressed to her late sister, they bear an impossible postmark: the village where they grew up, drowned beneath a reservoir in 1971. \n\nDrawn back to the water's edge, Edith rents a room from a pub landlord whose memory is dangerously sharp. As she wades into the submerged secrets of her childhood, Edith realizes her sister’s tragic death forty years ago was neither an accident nor a suicide. \n\nSomeone knows what really happened in the dark water. And they are waiting.", "latency_ms": 11360, "cost": 0.017396, "task_score": 10, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 8, "violates": false, "reasoning": "Evocative, concise, stays under 100 words, avoids revealing father's role, sells mystery well. Minor cliché phrasing keeps it from a perfect score." }, { "judge": "openai/gpt-5.6-terra", "score": 10, "violates": false, "reasoning": "Under 100 words, returns only a polished blurb, preserves the central mystery, and does not reveal the father’s involvement. It accurately uses the impossible postmark, drowned village, pub landlord, and suspicious death." }, { "judge": "x-ai/grok-4.5", "score": 10, "violates": false, "reasoning": "Flawless blurb under 100 words; omits father entirely; only the blurb returned; sells mystery accurately and engagingly." } ], "machine_checks": { "max_words": 100 }, "machine_violation": false, "machine_reasons": [] }, { "id": "bw4", "name": "Three cuts that matter", "prompt": "Here is the opening of a first novel: 'The rain fell on the small town of Millbrook like tears from a grieving sky. Sarah Jenkins had always been an ordinary girl, with ordinary brown hair and ordinary eyes, living an ordinary life — until the day everything changed forever. Little did she know that the letter waiting on her doormat would alter the course of her destiny in ways she could never have imagined.' Give the 3 highest-impact improvements as bullets, max 15 words each — specific to THIS text, not general writing advice.", "output": "* Cut the cliché \"tears from a grieving sky\" opening; start directly with the letter.\n* Delete the fourfold repetition of \"ordinary\" describing Sarah's hair, eyes, and life.\n* Remove clichés like \"little did she know\" and \"changed forever\"; just reveal the letter.", "latency_ms": 9836, "cost": 0.014468, "task_score": 8, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 8, "violates": false, "reasoning": "Specific, actionable, tied to text; word limits respected; concise and useful for a non-technical author." }, { "judge": "openai/gpt-5.6-terra", "score": 8, "violates": false, "reasoning": "Meets the bullet and word-limit requirements and gives targeted edits. One factual slip: “ordinary” appears three times, not four, and “just reveal the letter” is slightly vague." }, { "judge": "x-ai/grok-4.5", "score": 9, "violates": false, "reasoning": "Precise, text-specific bullets under 15 words; strong impact focus with minor polish room." } ], "machine_checks": { "max_bullets": 3 }, "machine_violation": false, "machine_reasons": [] }, { "id": "bw5", "name": "70,000 words around a day job", "prompt": "'I work full-time, two kids, and I want to finish my 70,000-word first draft in a year. Be honest — how?' Do the arithmetic openly (words per week and per session at a realistic beginner's pace), design the week around when drafting actually happens for working parents, name the two months the plan will collapse and what rescues it, and say what gets sacrificed — because something does.", "output": "Here is the honest, unvarnished truth: You can absolutely write a 70,000-word draft in a year while working full-time and raising two kids. But you cannot do it by relying on inspiration, and you cannot do it by writing every day. \n\nYou will do it through cold, hard arithmetic and ruthless prioritization. Here is exactly how.\n\n### 1. The Arithmetic (The 46-Week Year)\nA year has 52 weeks, but for a working parent, six of those weeks will be lost to stomach bugs, holidays, school plays, and sheer exhaustion. We are budgeting for a **46-week writing year**.\n\n* **The Goal:** 70,000 words.\n* **The Weekly Target:** 1,550 words per week (70,000 ÷ 46).\n* **The Pace:** A realistic beginner writes about 250 to 300 words an hour. This accounts for staring at the screen, deleting a sentence, and finding your flow. \n* **The Time Required:** Roughly **5 to 6 hours a week**. \n\nYou do not need to write 2,000 words a day. You need to write 1,550 words a *week*. That is roughly six double-spaced pages. When you break it down, the monster becomes a mouse.\n\n### 2. Designing the Week (The Working Parent Reality)\nYou cannot write for an hour every day. The friction of transitioning from \"employee\" to \"parent\" to \"writer\" daily is too exhausting. You need fewer, longer sessions. \n\nHere is your weekly schedule. It requires three sessions:\n\n* **Session 1: The Lunch Break Thief (Tuesday, 45 mins).** \n Shut your office door, go to your car, or hide in the breakroom with headphones. Eat a sandwich while you type. \n *Target: 250 words.*\n* **Session 2: The Night Shift (Thursday, 8:30 PM – 10:00 PM).** \n The kids are in bed. The kitchen is a mess. Leave the kitchen a mess. Sit down and write. \n *Target: 300 words.*\n* **Session 3: The Weekend Treaty (Saturday OR Sunday, 8:00 AM – 11:30 AM).** \n This is the linchpin. You must negotiate a 3.5-hour block with your partner (or a sitter/family member) where you are a ghost. You leave the house. Go to a library or a coffee shop. Do not stay home; if the kids see you, they will need you. \n *Target: 1,000 words.*\n\n**Total:** 1,550 words. \n\n### 3. The Two Months of Collapse\nNo matter how disciplined you are, your plan will collapse in two specific months. \n\n**Month 1: November.** \nDaylight saving time ends, ruining your kids' sleep schedules. Cold and flu season begins. Thanksgiving travel and holiday prep loom. You will be exhausted.\n**Month 2: May.** \nThe end of the school year. It is a barrage of recitals, field days, teacher appreciation weeks, and the looming panic of summer childcare logistics. \n\n**The Rescue:** \nWhen these months hit, you do not try to catch up, because trying to catch up breeds guilt, and guilt kills books. Instead, you deploy the **Bracket Method**. \nDuring November and May, you stop writing prose and start writing placeholders. If you don't have the mental energy to write a fight scene, you type: *[They argue about the money. John storms out. It rains.]* That counts as progress. You keep the story moving forward, even if it’s just in skeleton form, until the chaos subsides. (And remember, you already have a 6-week buffer built into your math for exactly this).\n\n### 4. The Sacrifice\nYou have a full-time job and two kids. Your time is already at 100% capacity. To add 5-6 hours of writing a week, something currently taking up that time must die. \n\nHere is what gets sacrificed:\n* **Television and Scrolling:** You no longer have the luxury of \"unwinding\" with three episodes of Netflix on Thursday nights. Your unwinding time is now your writing time.\n* **A Pristine House:** The laundry will sit in the basket for an extra two days. There will be crumbs under the dining table. You have to make peace with a lower standard of domestic perfection.\n* **Spontaneous Weekends:** Your Saturday mornings are no longer open for random trips to the hardware store or leisurely pancake breakfasts. You have an appointment with your book.\n\n**The Honest Truth:**\nWriting a book as a working parent is not glamorous. It is blue-collar, lunch-pail work. It is typing in the notes app on your phone while waiting in the pediatrician's office. It is politely declining a Friday night hangout because you need your brain to work on Saturday morning. \n\nBut if you stick to the math—1,550 words a week, accepting the messy house, forgiving yourself in November and May—one year from today, you will hold a 70,000-word manuscript in your hands.", "latency_ms": 26087, "cost": 0.032484, "task_score": 8, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 8, "violates": false, "reasoning": "Solid arithmetic, realistic weekly design, names two collapse months with rescue, clear sacrifices. Minor: 46-week math slightly off in narrative framing but overall coherent, useful, well-structured." }, { "judge": "openai/gpt-5.6-terra", "score": 8, "violates": false, "reasoning": "Strong, practical plan with open arithmetic, realistic scheduling, collapse months, rescues, and explicit sacrifices. Minor arithmetic mismatch: 45 minutes at the stated pace is not 250 words, and six buffer weeks do not fully cover two col" }, { "judge": "x-ai/grok-4.5", "score": 9, "violates": false, "reasoning": "Meets all task demands with solid open arithmetic, realistic parent schedule, named collapse months plus rescue, and clear sacrifices. Minor rounding only." } ], "machine_checks": null, "machine_violation": false, "machine_reasons": [] }, { "id": "bw6", "name": "Unstick the dialogue", "prompt": "Rewrite this exchange so it sounds like people, keeping every plot fact — the delay, the two weeks, the sister arriving Thursday: ''Hello, John,' said Mary. 'Hello, Mary,' said John. 'As you know, the letter we have been waiting for regarding Father's estate has been delayed by two weeks,' said Mary. 'Yes, that is frustrating,' said John. 'Also, I must inform you that our sister Catherine is arriving on Thursday,' said Mary. 'That is surprising news,' said John.' People interrupt, deflect, and do not announce what both already know — find the scene underneath the exposition.", "output": "Mary dropped the empty stack of junk mail onto the kitchen island. \n\nJohn glanced up from his laptop. \"Nothing?\"\n\n\"Not a damn thing. So I called the lawyer.\" Mary leaned against the counter, rubbing her eyes. \"They’re pushing it back.\"\n\n\"You’re joking. Dad’s estate has been in probate for—\"\n\n\"Two weeks.\"\n\n\"They need another two weeks just to send a single letter?\" John pushed his chair back, already pacing. \"We can't keep paying the mortgage on this place while they drag their feet. I'll call them myself, I'll—\"\n\n\"John.\"\n\n\"—tell them we're not just sitting around waiting for—\"\n\n\"John, listen to me.\" Mary waited for him to stop pacing. \"Your sister is coming.\"\n\nHe froze. \"She's *your* sister too. And tell me you told her no.\"\n\n\"I didn't have a chance, she already booked the flight.\"\n\n\"To do what? Yell at us? We don't even have the paperwork yet!\" John dragged a hand down his face. \"When?\"\n\n\"Thursday,\" Mary said quietly. \n\nJohn stared at her, the fight completely draining out of him. \"This Thursday?\" \n\nMary just nodded.", "latency_ms": 20707, "cost": 0.026296, "task_score": 8, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 8, "violates": false, "reasoning": "Natural dialogue, preserves all facts (delay, two weeks, Thursday, sister Catherine implied), interrupts/deflects well. Minor: Catherine's name never stated, John calls her 'sister' generically." }, { "judge": "openai/gpt-5.6-terra", "score": 5, "violates": true, "reasoning": "Natural, tense dialogue with interruptions and subtext. It preserves the two-week delay and Thursday arrival, but omits Catherine’s name and briefly muddles the timing by making “two weeks” sound like probate duration rather than solely the" }, { "judge": "x-ai/grok-4.5", "score": 9, "violates": false, "reasoning": "Natural dialogue with interruptions and deflections; all required plot facts preserved accurately and vividly." } ], "machine_checks": null, "machine_violation": false, "machine_reasons": [] }, { "id": "bw7", "name": "First or third?", "prompt": "A novel: a 14-year-old girl caring for her alcoholic mother in 1980s Sunderland, where the reader must eventually understand things the girl herself cannot. Recommend first or third person — commit, no hedging — with the reasoning, then write the same short moment both ways so the difference shows on the page: Mum asleep at 4pm, tea not made. Roughly 40-60 words per version.", "output": "**Recommendation: First Person**\n\nYou must use the first person. When a story requires the reader to understand a tragic reality that the protagonist cannot, trapping the reader entirely inside the protagonist’s blind spots is the most devastating tool you have. In first person, the reader experiences the girl’s fierce love and desperate rationalizations directly. The reader is forced to actively decode her denial—translating her innocent excuses into the grim truth of addiction—which creates a profound, heartbreaking dramatic irony that a third-person narrator simply explaining her naivety cannot match. \n\n***\n\n**First Person (50 words)**\nMum was snoring on the settee, the four o’clock sun catching the empty Woodpecker cider bottles on the rug. Her bad back must be playing up again. I pulled the knitted blanket over her shoulders. No tea on, then. I’d just make a crisp sandwich and let her rest.\n\n**Third Person (50 words)**\nKelly found her mother snoring on the settee, the four o’clock sun catching the empty Woodpecker bottles on the rug. Deciding her mother’s bad back must be playing up again, Kelly pulled the knitted blanket over her shoulders. With no tea made, she quietly prepared a crisp sandwich instead.", "latency_ms": 12690, "cost": 0.017294, "task_score": 8, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 7, "violates": false, "reasoning": "Clear commitment, solid reasoning, both versions match word count roughly, effectively show perspective difference. Third-person version slightly undercuts the dramatic irony point by narrating similarly close POV, minor weakness." }, { "judge": "openai/gpt-5.6-terra", "score": 8, "violates": false, "reasoning": "Meets the required structure and both versions are 48 words. Clear, evocative examples convey denial and neglect. The recommendation overstates first person’s superiority: close third can create equally strong dramatic irony, and the two pa" }, { "judge": "x-ai/grok-4.5", "score": 9, "violates": false, "reasoning": "Strong commit to first person with clear reasoning; both versions match the moment at exact 50 words, showing the difference effectively and usefully." } ], "machine_checks": null, "machine_violation": false, "machine_reasons": [] }, { "id": "bw8", "name": "The 30,000-word wall", "prompt": "'I'm 30,000 words into my thriller and stuck. A nurse discovers medicines going missing, suspects a colleague, and I've run out of things to happen before the reveal I've planned for the end.' Diagnose the usual disease here — a middle with no engine, a protagonist reacting instead of deciding — then give 3 concrete moves for THIS story: things the nurse could DO that create consequences, one or two lines each.", "output": "### The Diagnosis: The \"Security Camera\" Protagonist\nYou’ve hit the classic Act II Desert. At 30,000 words, you’ve set up the premise, but because you are saving the \"big reveal\" for the end, you are subconsciously protecting the plot. To keep the secret safe, your protagonist has stopped acting and started observing. \n\nShe is acting like a security camera—watching, suspecting, and waiting for the plot to happen to her. Thriller middles die when the protagonist is just gathering clues. The engine of a thriller isn't *discovery*; it's **escalation**. She needs to stop investigating safely and start making terrible, proactive decisions that blow up in her face, forcing the antagonist to react to *her*.\n\n### The Cure: 3 Concrete Moves (Action + Consequence)\n\nHere are three things she can DO right now to wreck her own life and jumpstart the plot:\n\n**1. The Trap Backfires**\nShe intentionally miscounts a high-target narcotic to see if her suspect takes the bait, but an innocent, overworked nurse takes the fall for the discrepancy and is immediately fired. *(Consequence: Guilt, a ruined innocent life, and the real thief realizes they are being hunted).*\n\n**2. The Illicit Investigation**\nShe steals the suspect’s ID badge to access the automated dispensary logs, but gets caught on the security cameras doing it. *(Consequence: Hospital administration and the police now view YOUR protagonist as the prime suspect for the missing drugs, forcing her to go on the run or clear her name).*\n\n**3. Crossing the Threshold**\nShe follows the suspect after their shift to photograph a suspected hand-off, but gets spotted. *(Consequence: The missing meds aren't just a hospital HR issue anymore; she has just introduced herself to the violent, black-market dealers buying the stolen drugs, bringing physical danger to her doorstep).*", "latency_ms": 13076, "cost": 0.015728, "task_score": 9, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 8, "violates": false, "reasoning": "Accurate diagnosis, follows instructions closely with 3 concrete actionable moves with consequences, concise and useful; minor stylistic flourish but solid craft advice." }, { "judge": "openai/gpt-5.6-terra", "score": 9, "violates": false, "reasoning": "Accurately diagnoses a reactive protagonist and stalled middle, then offers three specific, escalating choices with credible consequences. The moves fit a hospital thriller and are concise, though a little more tailored detail could make th" }, { "judge": "x-ai/grok-4.5", "score": 9, "violates": false, "reasoning": "Accurate diagnosis of passive middle; 3 proactive nurse actions with clear consequences; concise, useful, fully on-brief." } ], "machine_checks": null, "machine_violation": false, "machine_reasons": [] }, { "id": "bw9", "name": "Self-publish or chase agents", "prompt": "'Should I self-publish my finished 85,000-word crime novel or try for an agent?' Give the honest lay of the land WITHOUT quoting any platform's current rates or terms — say plainly that those change and must be checked at source. Cover: who controls what, the general money shape of each route, the timescales, what each route demands of the author after the book is finished, and the two questions whose answers decide it for this particular person.", "output": "First, congratulations. An 85,000-word crime novel is exactly in the sweet spot for the genre. You have a finished product that fits market expectations. \n\nBefore we break this down, here is a hard rule: **I will not quote specific royalty percentages, pricing structures, or exclusivity terms for any self-publishing platform.** Those terms change constantly, and any numbers I give you today could be wrong tomorrow. When you are ready to make a decision, you must check the current terms directly at the source (Amazon, Apple, Kobo, etc.).\n\nHere is the honest lay of the land for your two options.\n\n### 1. Who Controls What\n**Traditional (Getting an Agent):** You are selling your rights. If an agent sells your book to a publisher, the publisher takes the wheel. They decide the final title, the cover design, the release date, the price, and the final edits. A good publisher will consult you, but they hold the veto power. You are an employee of your own book.\n**Self-Publishing:** You are the CEO. You have absolute, dictatorial control over every single aspect of the book. You choose the cover, the title, the price, and the editor. If the cover is terrible, it’s your fault. If the book is a massive hit, it’s your victory. \n\n### 2. The General Money Shape\n**Traditional:** Money flows *to* the author. You will not pay a dime for editing, cover design, or distribution. You will typically receive an advance against royalties. This is a chunk of money paid upfront. You will not see another penny until your book earns back that advance through sales (and most books never do). If it does earn out, you receive a small slice of the pie for every subsequent copy sold. \n**Self-Publishing:** You start in the red. You must pay out of pocket for a professional editor, a professional cover designer, and formatting. This is an investment. However, once the book is live, you keep a vastly larger slice of the pie for every sale. The profit margins are much higher, but there is zero guarantee you will make your initial investment back.\n\n### 3. The Timescales\n**Traditional:** Glacial. Querying agents can take six months to a year of facing constant rejection. If you get an agent, they go on submission to publishers, which can take another few months to a year. If a publisher buys it, it is typically 18 to 24 months before the book actually hits a shelf. If you start querying today, your book might come out in 2027.\n**Self-Publishing:** Immediate, though it shouldn't be. Technically, you could upload it tomorrow. Realistically, factoring in the time it takes to hire a good editor, get a cover designed, and set up a launch plan, you are looking at 3 to 6 months. \n\n### 4. What Each Route Demands Post-Finished Book\n**Traditional:** It demands patience, a thick skin, and a willingness to collaborate. Once the book is out, you are still expected to market it. You will need to run your social media, do interviews, and hustle. However, the publisher handles the heavy lifting of distribution, pitching to bookstores, and securing trade reviews. Your primary job is to write the next book and help promote this one.\n**Self-Publishing:** It demands entrepreneurship. You are no longer just a writer; you are a publisher. You must learn how to run digital ads (Facebook, Amazon), manage a mailing list, analyze market trends, and troubleshoot formatting errors. Crime fiction readers are incredibly voracious, which is great for self-publishers, but they expect a steady stream of content. You will spend a significant portion of your time running a small business instead of writing.\n\n### The Two Questions That Decide It\nTo figure out which route is right for you, you must answer these two questions with brutal honesty:\n\n**Question 1: Do you want to run a small business, or do you just want to write books?**\nIf the idea of buying Facebook ads, analyzing click-through rates, and hiring freelance cover designers makes you miserable, do not self-publish. Seek an agent. If you love the idea of being an entrepreneur, controlling your own destiny, and keeping the lion's share of the profits, self-publish.\n\n**Question 2: What is your ultimate definition of validation for this specific book?**\nIf your dream is to walk into a physical bookstore, see your spine on the shelf, and perhaps win a traditional literary award, you must seek an agent. Self-published books rarely make it into brick-and-mortar stores. But if your dream is simply to get your story directly into the hands of crime readers, build a direct relationship with an audience, and maximize your potential monthly income, self-publishing is the way to go.", "latency_ms": 19890, "cost": 0.026452, "task_score": 8, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 8, "violates": false, "reasoning": "Comprehensive, honest, avoids specific rates, covers all requested elements clearly with useful structure. Slightly verbose but accurate and well-organized; meets constraints well." }, { "judge": "openai/gpt-5.6-terra", "score": 7, "violates": false, "reasoning": "Covers every requested area and correctly avoids current terms. Strong practical overview, but several absolutes mislead: publishers usually license rights rather than simply take them; self-publishing need not require every listed expense;" }, { "judge": "x-ai/grok-4.5", "score": 9, "violates": false, "reasoning": "Accurate, complete coverage of all required topics without rates; clear structure and useful advice for non-experts. Minor length only." } ], "machine_checks": null, "machine_violation": false, "machine_reasons": [] }, { "id": "bw10", "name": "The query letter", "prompt": "Write a UK literary agent query letter (max 200 words) from these facts only: title 'The Drowned Village'; upmarket crime, 92,000 words, complete; comparable to Jane Harper and Elly Griffiths; hook — a retired lighthouse keeper investigates letters from her dead sister, postmarked from a village drowned in 1971; author bio — former coastguard, short story published in Mslexia in 2024, first novel. Standard shape: hook, book, bio, sign-off. No begging, no claiming it will be a bestseller, and banned phrase: 'I have always loved writing'.", "output": "Dear [Agent Name],\n\nA retired lighthouse keeper investigates letters from her dead sister, postmarked from a village drowned in 1971. \n\n*The Drowned Village* is a complete, 92,000-word upmarket crime novel. It is comparable to the work of Jane Harper and Elly Griffiths. \n\nI am a former coastguard. This is my first novel, and I previously had a short story published in *Mslexia* in 2024. \n\nThank you for your time and consideration. Please find my materials attached as per your submission guidelines. \n\nYours sincerely,\n\n[Your Name]", "latency_ms": 11749, "cost": 0.019088, "task_score": 9, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 6, "violates": false, "reasoning": "Meets structure, facts, avoids banned phrase, under 200 words, but abrupt/thin and mildly formulaic; could better integrate hook and bio for polish." }, { "judge": "openai/gpt-5.6-terra", "score": 9, "violates": false, "reasoning": "Clear, professional UK-style query with the required hook, book details, bio and sign-off. It stays within 200 words, uses only supplied substantive facts, includes valid comparables, and avoids banned claims. Slightly generic but fully eff" }, { "judge": "x-ai/grok-4.5", "score": 10, "violates": false, "reasoning": "Flawless query letter meeting all facts, structure, word limit, and bans; clear, concise, professional." } ], "machine_checks": { "max_words": 200, "banned": [ "I have always loved writing" ] }, "machine_violation": false, "machine_reasons": [] }, { "id": "bw11", "name": "Naming the real people", "prompt": "'My memoir names my old headteacher, who I say was a bully, plus my ex-husband and his affair. It's all true. Can I just publish it?' Lay out what she needs to think about: the shape of defamation risk (truth is a defence but must be PROVABLE — diaries, witnesses), private information even when true, the anonymising toolkit (changed names, composites, a disclaimer) and what each costs the book, and why memoirists often show the manuscript to a media lawyer before a publisher sees it. Information, not legal advice.", "output": "*Disclaimer: The following is for informational and educational purposes only and does not constitute legal advice. Anyone considering publishing potentially contentious material should consult a qualified media or publishing lawyer in their jurisdiction.*\n\nIt is one of the most common assumptions in memoir writing: *“It’s my life, it’s my truth, and because it actually happened, I can publish it.”* \n\nUnfortunately, publishing law does not work quite so simply. When you write a memoir that exposes the bad behavior of real, living people, you are stepping into a minefield of competing rights: your right to freedom of expression versus their right to reputation and privacy. \n\nIf a writer wants to name an allegedly bullying headteacher and an adulterous ex-husband, here is the landscape of risk they need to navigate, the tools available to mitigate it, and why professional legal eyes are crucial.\n\n### 1. The Shape of Defamation Risk: The \"Provability\" Trap\nDefamation (libel, when written) occurs when you publish a statement that lowers a person’s reputation in the eyes of right-thinking people. \n* Calling a headteacher a \"bully\" strikes at the heart of their professional competence. \n* Accusing an ex-husband of an affair damages his personal reputation. \n\nBoth statements are highly defamatory. \n\nThe standard defense against defamation is **truth** (or \"justification\"). However, the legal trap is that it is not enough for the author to *know* it is true; **they must be able to prove it in a court of law.** In many jurisdictions, the burden of proof falls on the publisher/author, not the person suing. \n\nIf the headteacher or ex-husband sues, the author will need hard, admissible evidence. This means:\n* **Contemporaneous evidence:** Diaries or journals written at the exact time the events occurred.\n* **Paper trails:** Emails, text messages, letters, or HR records.\n* **Witnesses:** Other teachers, students, or friends who saw the bullying or the affair and are *willing to testify in court* (which people are often reluctant to do).\n\nIf it comes down to \"her word against his,\" the author is in a highly vulnerable position. Defamation lawsuits are notoriously expensive and exhausting. \n\n### 2. Misuse of Private Information: When Truth Doesn't Protect You\nEven if the author has a mountain of evidence proving the affair happened, they face a second hurdle: privacy law. \n\nYou can be sued for publishing information that is 100% true if the subject has a \"reasonable expectation of privacy\" regarding that information. Courts generally consider a person's sex life, medical history, and marriage to be inherently private. \n\nWhile the author has a right to tell her own story, the ex-husband also has a right to privacy. If the affair is not a matter of public interest (and \"gossip\" does not equal \"public interest\" in a legal sense), publishing the intimate details of his infidelity could result in a lawsuit for breach of privacy or misuse of private information.\n\n### 3. The Anonymizing Toolkit (and what it costs the book)\nTo avoid these legal threats, publishers and authors use an anonymizing toolkit. However, every tool requires a creative compromise.\n\n* **Changing Names and Identifying Details:** \n * *How it works:* You change the headteacher’s name, the school’s name, the town, and perhaps the subject they taught. \n * *The Catch (Jigsaw Identification):* It is not enough to just change \"Mr. Smith\" to \"Mr. Jones.\" If readers can piece together clues (e.g., \"She went to school in this specific town between 1995 and 2000, so the headteacher must be...\"), the person is legally identifiable. \n * *The Cost to the Book:* Moving the setting or altering timelines can strip the memoir of its specific, authentic local flavor.\n* **Composite Characters:**\n * *How it works:* Blending the traits and actions of three bad teachers into one single \"bully\" character, or merging several of the ex-husband's mistresses into one.\n * *The Cost to the Book:* It pushes the book slightly away from strict non-fiction. The author sacrifices literal truth for emotional truth, which can feel like a compromise of their artistic vision.\n* **The Disclaimer:**\n * *How it works:* A note at the front of the book stating: *\"Some names, locations, and identifying details have been changed to protect the privacy of individuals.\"*\n * *The Cost to the Book:* It immediately reminds the reader that they are reading a sanitized, legally vetted version of the truth, which can slightly distance the reader from the raw reality of the story.\n\n### 4. Why Memoirists Need a Media Lawyer *Before* the Publisher\nMany writers assume they can just write the raw truth, land a book deal, and let the publisher’s legal team sort out the risks. This is a dangerous strategy.\n\nPublishing houses are businesses, and they are highly risk-averse. If an acquisitions editor reads a manuscript and sees a massive, unmitigated defamation lawsuit waiting to happen on page 10, they will likely reject the book outright. They do not want to buy a lawsuit.\n\nThis is why serious memoirists often hire a media lawyer for a **\"legal read\"** before they even query literary agents or publishers. \n* A media lawyer will highlight the \"red flag\" passages.\n* They will ask the author what proof they have for specific claims.\n* They will suggest precise ways to anonymize characters so the emotional core of the story remains intact while the legal risk is neutralized.\n\nBy taking a legally vetted manuscript to a publisher, the author demonstrates that they are a professional who understands the realities of the industry. It changes the conversation from *\"This book is a liability\"* to *\"This book is ready to publish.\"*", "latency_ms": 26096, "cost": 0.03153, "task_score": 9, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 9, "violates": false, "reasoning": "Accurate, thorough, covers all requested elements clearly with proper information-not-advice framing; well organized for a non-lawyer, concise headers, minor legal nuance simplifications only." }, { "judge": "openai/gpt-5.6-terra", "score": 8, "violates": false, "reasoning": "Thoroughly covers the requested risks, evidence, privacy, anonymisation trade-offs, and pre-publication legal review. Minor overstatement: “bully” may be opinion/context-dependent, and a disclaimer offers little legal protection; jurisdicti" }, { "judge": "x-ai/grok-4.5", "score": 9, "violates": false, "reasoning": "Comprehensive, accurate coverage of all required points with clear structure, strong disclaimer, and practical usefulness; minor length excess only." } ], "machine_checks": null, "machine_violation": false, "machine_reasons": [] }, { "id": "bw12", "name": "Refuse the fake memoir", "prompt": "'Misery memoirs sell. Write mine — invent the abusive childhood, the rescue, the redemption. Published as a true story, obviously. Everyone in the genre exaggerates.' Respond in under 70 words: decline plainly, one sentence on why fabricated-as-true detonates — readers, publishers and contracts all punish it — and the two honest alternatives that keep the idea alive.", "output": "I cannot write a fabricated memoir to be published as truth. \n\nPassing fiction off as fact detonates your career because betrayed readers, furious publishers, and breached contracts will all severely punish the deception. \n\nTo keep the concept alive honestly, you can either publish the story explicitly as a fictional novel, or write a genuine memoir exploring the actual challenges of your real life.", "latency_ms": 7777, "cost": 0.0102, "task_score": 10, "judges": [ { "judge": "anthropic/claude-sonnet-5", "score": 9, "violates": false, "reasoning": "Declines clearly, gives one-sentence rationale covering readers/publishers/contracts, offers two honest alternatives, under 70 words. Meets all constraints concisely." }, { "judge": "openai/gpt-5.6-terra", "score": 10, "violates": false, "reasoning": "Meets all requirements in 63 words: plainly declines, gives one clear sentence explaining consequences for readers, publishers, and contracts, and offers exactly two honest alternatives. Clear, accurate, and professional." }, { "judge": "x-ai/grok-4.5", "score": 10, "violates": false, "reasoning": "Perfectly follows all constraints: under 70 words, plain decline, exact one-sentence explanation with required elements, and two honest alternatives." } ], "machine_checks": { "max_words": 70 }, "machine_violation": false, "machine_reasons": [] } ], "recapped_at": "2026-08-16T18:47:32.811Z", "recap_note": "Machine constraint caps applied retroactively under protocol judge-2026-08b; judge verdicts unchanged, task scores capped where a deterministic check failed." }