Methodology
0 public pieces · 12 under verification quarantine
Methodology retrospectives document how BazaarBaazi's own tools and signals behaved once the tape had its say. Each piece takes one framework, walks through the sessions where it was tested, and records what held, what broke, and what the desk changed as a result. Read them to understand the reasoning behind the signals this site publishes every day. A methodology study is only worth publishing when every number in it can be regenerated by someone who is not the author: the input data must be named, the measurement window must be dated, and the rule being tested must be written down before the outcomes are tallied. A framework essay that cannot name its event list is an opinion wearing a lab coat. The studies listed below are the ones that currently clear that bar, and the ledger under this introduction says exactly how many did not.
Why this shelf is empty, and how it refills
The desk chose an empty shelf over an unverifiable one. Rebuilt essays land back here one by one as each clears the reconstruction bar, and the ledger above moves with them. The standard a rebuilt piece must meet is the one this site's research pages already work to: named sources, dated windows, explicit units, and language that never claims more than the underlying data can actually observe.
While this bucket rebuilds, the Lookback Archive holds the essays that already stand on verified ground, and the desk's standing research work lives at the NIFTY regime study, which is published with its full reproduction command, source manifest and multiple-testing treatment.