SkrrrrLabs

Trade-level flow as evidence, not instruction.

More granular data is not automatically better evidence. The question is whether trade-level information remains useful outside the period in which it was discovered.

SourceBinance USDT-M Futures
Data classTrade-level aggTrades
Source archives processed1,518
Matched events1,852
Deployment stateNot deployed
01

Research question

Does trade-level execution flow add stable information after market regime, chronology and out-of-sample evaluation are held constant?

02

Data boundary

The study uses independently maintained public exchange archives. Events are matched chronologically and divided into development, validation and sealed evaluation periods.

03

Current finding

The additional detail improved some development-period descriptions, but the improvement did not persist strongly enough in sealed evaluation to support deployment. The result remains a research note rather than a trading system.

04

Why the negative result matters

A larger dataset can make a weak relationship look precise. Rejecting unstable improvements is part of the research process and prevents data volume from being mistaken for evidence.

05

What remains private

Feature definitions, thresholds, event labels, instrument-level records and executable decision logic are not published.