Our calls, on the record.
Anyone can claim they saw it coming. Every month we freeze the REQ Score™ for every drug we score and publish the file's fingerprint here, before anyone knows how it turns out. Twelve months later we publish the file itself. If one character had changed, the fingerprint would not match.
How it works
- Freeze. At least every 28 days the score build saves every drug's crowding risk, supply risk, value score and band, using only FDA and CMS data public that day.
- Fingerprint. We publish the file's SHA-256 fingerprint below. A fingerprint is a 64-character code computed from every byte of the file; change anything and the code changes.
- Reveal. On the reveal date we publish the file. Check it yourself: on a Mac or Linux run
shasum -a 256 <file>; on Windows runcertutil -hashfile <file> SHA256. The result must equal the fingerprint posted here. - Grade. We grade every frozen file against what actually happened, the same way as our back-tests, and post the result, good or bad.
Frozen predictions
Loading the ledger…
What the back-tests already show
Before the first freeze, every part of the REQ Score™ was tested on past dates it was not built on. On three past dates, drugs we rated high crowding risk gained a new maker far more often than drugs rated low (ranking power 0.79 to 0.82, where 0.5 is a coin flip), and drugs rated high supply risk lost a maker 2.6 to 2.9 times as often. The frozen files above are the forward test: the part nobody can tune after the fact.
The REQ Score™ is an informational indicator, not investment, legal or regulatory advice. Shortage prediction is not part of these files.