Forecast Ledgerinterventional psychiatry and psychedelic medicineQuoted, dated predictions scored against public outcomes under one written method. Every score links to the words, the rule and the evidence.

Not admitted · Vague or promotional

Rules 1.1.0 · as of 14 Sep 2026

Quote

Vague or promotional

“any decent machine learning engineer can back-solve DRG multipliers based on claims data. So, good luck keeping that Private and proprietary.”
· We Made Inpatient Psychiatric Treatment Suck Less with this One Weird Trick · substack post

Footnote 2 of the hospital payment post, about the DRG multipliers that third-party payers treat as proprietary. The author says any decent machine learning engineer can back-solve them from claims data, so payers will not manage to keep them private.

No entity, observable event, benchmark or threshold can be written from the text and its immediate context. Slogans ('game changer', 'once-in-a-lifetime opportunity', 'favourite stock') and quantifiers with no countable set ('more', 'first', 'all') without a defined set.