PROMOTE: Anytime-Valid Promotion Tests for Adaptive Financial Research Agents
What it is
This paper introduces PROMOTE, a statistical protocol that assigns predictable testing wealth to trading rule candidates at birth and accumulates cost-aware e-processes to control false discovery rates. It addresses the issue of adaptive selection in financial research agents by using a weighted Ville argument to bound the probability of promoting rules with nonpositive conditional mean, even when candidates are born at data-dependent times and monitored indefinitely.
The problem it solves
Financial research agents generate and monitor trading rules faster than humans can document the selection process, leading to high rates of false positives due to adaptive selection and multiple testing.
What you could use it for
- Controlling false discovery rates in automated trading strategy research
- Validating adaptive financial research agents that generate and monitor multiple trading rules
- Auditing the statistical integrity of backtested trading strategies with adaptive selection
Why it matters
Prevents the publication of spurious trading edges by rigorously controlling for the multiple testing and adaptive selection inherent in AI-driven financial research, thereby increasing the reliability of automated strategy discovery.
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.