$ styleprofile score examples/writer/sharpening.md examples/writer/old-maps.md '<tmp>/writer.json'
[exit 0]
--- stdout
STYLE COMPARISON   2 documents, 2 chunks, 1,277 words   vs writer.json (7 chunks)

By document   furthest from the reference first
  2 documents: 2 close
  document                      words  verdict             Delta  LLM-likeness
  examples/writer/old-maps.md     640  close                0.52  like reference
  examples/…ter/sharpening.md     637  close                0.47  like reference

Across 2 documents: close   Delta 0.49
  Lower is closer. The reference's own held-out writing at these lengths (637-640 words per chunk) scores 0.68 typically
  (a mean over 2 chunks, up to 0.86).
  Pooled over every document from here on; the table above has each one's own.

LLM-likeness: like the reference   0.10
  The reference's own writing at these lengths (637-640 words per chunk) scores 0.20 typically (a mean over 2 chunks, up
  to 0.73); the LLM drafts 14.23.

By area   Delta ÷ the top of the reference's usual range in each area; bars on a log scale
  close up to 1x, somewhat different to 1.5x, clearly different to 2x, very different above
  Vocabulary                   0.68x  ███░░░░░░░░░░░░░░░░░  close
  Function words               0.62x  ███░░░░░░░░░░░░░░░░░  close
  Voice and stance             0.62x  ███░░░░░░░░░░░░░░░░░  close
  Punctuation                  0.57x  ███░░░░░░░░░░░░░░░░░  close
  Rhythm                       0.38x  ██░░░░░░░░░░░░░░░░░░  close
  Sentence shape               0.26x  █░░░░░░░░░░░░░░░░░░░  close
  Formatting                   0.00x  ░░░░░░░░░░░░░░░░░░░░  close

Biggest differences   (each ▲ or ▼ is one standard deviation, up to 3)
                                                   this text   reference
  Parentheses                                        3.1 /1k     1.3 /1k   ▲
  Hedges (maybe, I think)                            7.0 /1k     3.6 /1k   ▲
  Semicolons                                         1.6 /1k     2.9 /1k   ▼
  "did"                                              3.9 /1k     1.5 /1k   ▲
  "he"                                               7.8 /1k     4.0 /1k   ▲
  "was"                                              9.4 /1k     5.6 /1k   ▲
  "from"                                             0.8 /1k     2.7 /1k   ▼
  "some"                                             0.0 /1k     0.9 /1k   ▼

Pass --all to see every metric.
