| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 7 | | adverbTagCount | 0 | | adverbTags | (empty) | | dialogueSentences | 41 | | tagDensity | 0.171 | | leniency | 0.341 | | rawRatio | 0 | | effectiveRatio | 0 | |
| 96.49% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 1425 | | totalAiIsmAdverbs | 1 | | found | | | highlights | | |
| 100.00% | AI-ism character names | Target: 0 AI-default names (17 tracked, −20% each) | | codexExemptions | (empty) | | found | (empty) | |
| 100.00% | AI-ism location names | Target: 0 AI-default location names (33 tracked, −20% each) | | codexExemptions | (empty) | | found | (empty) | |
| 78.95% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 1425 | | totalAiIsms | 6 | | found | | | highlights | | 0 | "perfect" | | 1 | "warmth" | | 2 | "etched" | | 3 | "magnetic" | | 4 | "pulsed" | | 5 | "mechanical" |
| |
| 100.00% | Cliché density | Target: ≤1 cliche(s) per 800-word window | | totalCliches | 0 | | maxInWindow | 0 | | found | (empty) | | highlights | (empty) | |
| 100.00% | Emotion telling (show vs. tell) | Target: ≤3% sentences with emotion telling | | emotionTells | 0 | | narrationSentences | 126 | | matches | (empty) | |
| 100.00% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 0 | | hedgeCount | 0 | | narrationSentences | 126 | | filterMatches | (empty) | | hedgeMatches | (empty) | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 160 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 31 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 0 | | markdownWords | 0 | | totalWords | 1420 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 7 | | unquotedAttributions | 0 | | matches | (empty) | |
| 100.00% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 24 | | wordCount | 1105 | | uniqueNames | 9 | | maxNameDensity | 0.9 | | worstName | "Quinn" | | maxWindowNameDensity | 1.5 | | worstWindowName | "Quinn" | | discoveredNames | | Harlow | 1 | | Quinn | 10 | | Victorian | 1 | | Fellowes | 4 | | Morris | 2 | | Hackney | 1 | | TfL | 1 | | Bovril | 1 | | Warm | 3 |
| | persons | | 0 | "Harlow" | | 1 | "Quinn" | | 2 | "Fellowes" | | 3 | "Morris" | | 4 | "Warm" |
| | places | (empty) | | globalScore | 1 | | windowScore | 1 | |
| 83.33% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 75 | | glossingSentenceCount | 2 | | matches | | 0 | "something between a maze and a sigil, the kind" | | 1 | "looked like the tattoo on this dead man's" |
| |
| 59.15% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 2 | | per1kWords | 1.408 | | wordCount | 1420 | | matches | | 0 | "Not in one, not near one, but precisely between two of them" | | 1 | "not near one, but precisely between two of them" |
| |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 0 | | totalSentences | 160 | | matches | (empty) | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 67 | | mean | 21.19 | | std | 19.33 | | cv | 0.912 | | sampleLengths | | 0 | 21 | | 1 | 53 | | 2 | 43 | | 3 | 61 | | 4 | 3 | | 5 | 21 | | 6 | 1 | | 7 | 2 | | 8 | 41 | | 9 | 3 | | 10 | 15 | | 11 | 14 | | 12 | 25 | | 13 | 55 | | 14 | 8 | | 15 | 51 | | 16 | 1 | | 17 | 1 | | 18 | 8 | | 19 | 25 | | 20 | 54 | | 21 | 5 | | 22 | 14 | | 23 | 31 | | 24 | 14 | | 25 | 66 | | 26 | 72 | | 27 | 18 | | 28 | 56 | | 29 | 5 | | 30 | 6 | | 31 | 3 | | 32 | 9 | | 33 | 46 | | 34 | 57 | | 35 | 8 | | 36 | 10 | | 37 | 1 | | 38 | 31 | | 39 | 9 | | 40 | 51 | | 41 | 25 | | 42 | 4 | | 43 | 6 | | 44 | 23 | | 45 | 49 | | 46 | 12 | | 47 | 14 | | 48 | 25 | | 49 | 1 |
| |
| 88.55% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 6 | | totalSentences | 126 | | matches | | 0 | "been replaced" | | 1 | "been shuttered" | | 2 | "been cooked" | | 3 | "was positioned" | | 4 | "been pulled" | | 5 | "was etched" |
| |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 1 | | totalVerbs | 176 | | matches | | |
| 100.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 0 | | semicolonCount | 0 | | flaggedSentences | 0 | | totalSentences | 160 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 1110 | | adjectiveStacks | 0 | | stackExamples | (empty) | | adverbCount | 35 | | adverbRatio | 0.03153153153153153 | | lyAdverbCount | 6 | | lyAdverbRatio | 0.005405405405405406 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 160 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 160 | | mean | 8.88 | | std | 7.04 | | cv | 0.794 | | sampleLengths | | 0 | 21 | | 1 | 15 | | 2 | 26 | | 3 | 6 | | 4 | 6 | | 5 | 17 | | 6 | 7 | | 7 | 2 | | 8 | 1 | | 9 | 16 | | 10 | 8 | | 11 | 27 | | 12 | 6 | | 13 | 1 | | 14 | 1 | | 15 | 6 | | 16 | 12 | | 17 | 3 | | 18 | 21 | | 19 | 1 | | 20 | 2 | | 21 | 5 | | 22 | 5 | | 23 | 17 | | 24 | 14 | | 25 | 3 | | 26 | 4 | | 27 | 11 | | 28 | 4 | | 29 | 3 | | 30 | 7 | | 31 | 25 | | 32 | 12 | | 33 | 14 | | 34 | 13 | | 35 | 16 | | 36 | 8 | | 37 | 20 | | 38 | 2 | | 39 | 5 | | 40 | 24 | | 41 | 1 | | 42 | 1 | | 43 | 8 | | 44 | 5 | | 45 | 20 | | 46 | 10 | | 47 | 5 | | 48 | 4 | | 49 | 17 |
| |
| 65.83% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 8 | | diversityRatio | 0.425 | | totalSentences | 160 | | uniqueOpeners | 68 | |
| 62.31% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 2 | | totalSentences | 107 | | matches | | 0 | "Then she stood and surveyed" | | 1 | "Away from the circles." |
| | ratio | 0.019 | |
| 100.00% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 31 | | totalSentences | 107 | | matches | | 0 | "He held his notebook like" | | 1 | "His shoes told a different" | | 2 | "She pulled a pen from" | | 3 | "She studied the man's hands." | | 4 | "She wanted to hear him" | | 5 | "He glanced at the circles." | | 6 | "She walked to the nearest" | | 7 | "She pressed her thumb to" | | 8 | "She pulled her thumb away" | | 9 | "His jaw worked sideways, the" | | 10 | "She'd missed it the first" | | 11 | "She pulled out her phone" | | 12 | "She paced the distance." | | 13 | "She checked the third, further" | | 14 | "She indicated the dust coating" | | 15 | "They both looked." | | 16 | "She ignored his tone." | | 17 | "She'd heard it before, from" | | 18 | "She was already here." | | 19 | "Her torch cut into the" |
| | ratio | 0.29 | |
| 100.00% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 76 | | totalSentences | 107 | | matches | | 0 | "The body lay face-down on" | | 1 | "Detective Harlow Quinn ducked under" | | 2 | "The abandoned station still held" | | 3 | "Someone had been here a" | | 4 | "DS Fellowes said, falling into" | | 5 | "He held his notebook like" | | 6 | "Quinn crouched beside the body" | | 7 | "The man wore a heavy" | | 8 | "His shoes told a different" | | 9 | "She pulled a pen from" | | 10 | "She studied the man's hands." | | 11 | "A callus on his right" | | 12 | "Fellowes flipped a page." | | 13 | "She wanted to hear him" | | 14 | "Quinn stood and looked at" | | 15 | "The forensics team had set" | | 16 | "The dust, the dead air," | | 17 | "The tile within each circle" | | 18 | "He glanced at the circles." | | 19 | "She walked to the nearest" |
| | ratio | 0.71 | |
| 0.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 0 | | totalSentences | 107 | | matches | (empty) | | ratio | 0 | |
| 0.00% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 42 | | technicalSentenceCount | 9 | | matches | | 0 | "The body lay face-down on the platform, arms outstretched like a man reaching for a departing train that would never come." | | 1 | "The man wore a heavy coat, expensive but threadbare at the cuffs, the kind of garment that had been fine six months ago and hadn't been replaced." | | 2 | "The tile within each circle had changed colour, darkened to a deep arterial red, as though the ceramic itself had been cooked from below." | | 3 | "A series of interlocking lines that formed something between a maze and a sigil, the kind of mark she'd seen once before in a file she'd never been able to clos…" | | 4 | "Nobody had listened, because the only evidence that didn't fit the stroke narrative was a symbol scratched into the concrete floor beside Morris's body." | | 5 | "The forensics team had laid down stepping plates, aluminium paths that hovered above the dust to preserve evidence." | | 6 | "That particular note of professional concern that meant you're going somewhere we can't follow." | | 7 | "The casing had gone green with verdigris, but the face was etched with markings that belonged to no cardinal system she recognised." | | 8 | "She looked at the compass needle, straining against its pivot, pointing at the wall like a dog scenting blood." |
| |
| 100.00% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 7 | | uselessAdditionCount | 0 | | matches | (empty) | |
| 100.00% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | | totalTags | 3 | | fancyCount | 0 | | fancyTags | (empty) | | dialogueSentences | 41 | | tagDensity | 0.073 | | leniency | 0.146 | | rawRatio | 0 | | effectiveRatio | 0 | |