21  Experiments

The running record of research results and how to reproduce them. This is distinct from research-platform/findings/, which holds polished, publication-style write-ups of results that have already stabilized — experiments here are the raw material those get built from, including ones that didn’t pan out (see journal/dead-ends.md).

Status: one experiment logged under this convention so far (below). Older results still live in research-platform/findings/ until they’re migrated or superseded.

21.1 Logged experiments

• 2026-08-03-factor-catalogue-empirical-results — provenance record (run id, config lookups, method) for the empirical fields in phd/factor-catalog.md, the per-factor, per-family catalog. Naive and HAC (Newey–West) t-stat fields are specified but pending computation.

21.2 Convention for a new experiment entry

Each experiment gets a dated folder, experiments/YYYY-MM-DD-short-name/, containing:

• README.md — what question (research-questions.md) this tests, what was run, and the result.
• the config / parameters used (or a link to the commit/run-id in SBFoundation if the run itself lives there — see Reproducibility & experiment infrastructure for what a run identity captures).
• a link to the journal entry that motivated it, and the entry that recorded the outcome.

21.3 Existing results (published-quality)

• asset-dependency-networks-econophysics-review.md
• edgar-pit-integrity.md

21.4 Reproducibility

Every experiment inherits the platform’s point-in-time and run-identity guarantees — see Reproducibility & experiment infrastructure for what “auditable and repeatable” means concretely here.