HeadroomDESK

Headroom Research

Headroom Research publishes the program's finished studies as they were filed: the question, what was registered before the test, the method, and the verdict. A null result gets the same page, the same care and the same prominence as any other.

Pending publication

A pre-registered forward test of BTC event-risk alerts closed on 2026-11-08. Protocol sha256 dee73a5628eb2afd4ac3d51d41e2158e6c36a33414a9cf8740a90d8807361ab0, recorded 2026-07-29 19:14 UTC. Close report sha256 b8d435fe9f5328771bb3c9c910cbdc2accfa03675e9b346711bed774a329f964. Publication of the results is pending a data-terms review.

Headroom Research, notice posted 2026-11-19.

Published

How a study is published

  • As filed. A release reports the filed, graded record. It adds no new computation, no re-rounding and no figure combined across studies; a question that needs one is a new study with its own page.
  • Registered first. Each release says what was fixed before the test, links the registration, and gives the sha256 of the locked page. A study registered internally before a public registry existed says so, with its dates; it is never presented as registered earlier than it was posted.
  • Every number with its sample and window. A number on a release page appears exactly as on the filed record, with its n and its data window in the same sentence, and the record's sha256 is printed on the page. An automatic check refuses any page that breaks this.
  • Misses as prominent as hits. An unfavourable result sits in the same block, in the same type, as a favourable one.
  • Fences disclosed. Each release names the data restrictions that applied and any event that touched them.
  • No calls. A verdict describes a test. None of it says what price will do next or is a reason to open or close a position.

Where the records live

This site is the home page of every release. Code and test files are on GitHub under the headroomdesk organisation and archived on Zenodo, which gives each archived version a DOI. Registrations are on OSF. The author on every record is "Headroom Research".

Errors found after publication are fixed in public in the corrections log.