Point-in-Time Data Auditing and a Negative BTC Perpetual Strategy Study

Summary

This study examines whether public Binance BTCUSDT perpetual-futures archives support valid point-in-time factor research. It distinguishes event, publication, and availability times, and separates proposal, audit, evaluation, and holdout stages. An initial requirement for gapless five-minute trade, mark, index, and open-interest data failed because the common record did not span the required history. The researchers revised the specification: open interest became optional because its publication time could not be verified, while trade, mark, index, and realized funding formed the core data set. The revised availability mask supported a chronological train, validation, and historical holdout split.

The deterministic auditor caught every tested violation among 80 frozen templates and rejected none of the legal templates. On ten null-signal paths, auditing reduced average false passes. Yet, with valid-candidate budgets matched, an adaptive agent only tied random search; superiority was not established. All evaluated historical holdout runs had positive information coefficients but negative net Sharpe under primary costs. These findings are specific to the audited archive, rules, and costs; they do not establish profitability or general agent performance.

Key ideas

• Point-in-time research requires checking when each observation became available, not just whether an archive contains it.
• The original four-stream completeness requirement failed, leading to a disclosed revision that made open interest optional.
• The auditor detected all tested violations and accepted all tested legal templates.
• With candidate budgets matched, adaptive search did not outperform random search.
• Historical holdout runs had positive information coefficients but negative net Sharpe under primary costs.

Tags

# 2608.25348


# Point-in-Time Audit Before Alpha: Public-Archive Availability and a Negative Matched-Budget Study on BTC Perpetual Futures









Public cryptocurrency archives may appear usable when files exist, although factor research requires observations available and executable at each decision time. We audit public Binance BTCUSDT USD-M perpetual-futures data using event, publication, and availability times and separate proposal from deterministic auditing, evaluation, and holdout access. An initial gapless five-minute requirement for trade, mark, index, and open interest failed: the longest unrepaired intersection was 304.5729166666667 days. A disclosed revision made trade, mark, index, and realized funding the core streams and made open interest optional because its publication time was unverified. The revised mask retained 727 complete UTC days and supported a 436/145/146-day train, validation, and historical-holdout split. On 80 frozen known-rule templates, the auditor detected 40/40 violations and rejected 0/40 legal templates. Across ten null-signal paths, full auditing reduced mean false passes from 0.2910 to 0.0625. Under matched valid-candidate budgets, the audited adaptive agent tied random search and did not establish superiority. In the one-time historical holdout, all evaluated runs had positive IC but negative net Sharpe under primary costs. We therefore report a scoped negative result rather than a profitability or agent-superiority claim.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.