Skip to main content

# Financial Backtesting

2 related articles

Subscribe to updates

arXiv Nov 5, 2026
Finance agents meet an open world: none of 150 sessions passes, and all 35 the simulated user approved fail a program check
FinCosmos opens all three uncertainties of financial agent evaluation at once: a simulated investor reveals a hidden strategy only when asked, the agent must source its own Web data and backtest under a data cutoff, and programs verify the delivery instead of matching a reference answer. Across 150 sessions with four coding agents and a tool-free model, none is accepted, and all 35 sessions the simulated user signed off fail at least one program veto.
FinCosmos opens all three uncertainties of financial agent evaluation at once: a simulated investor reveals a hidden strategy only when asked, the agent must source its own Web data and backtest under a data cutoff, and programs verify the delivery instead of matching a reference answer. Across 150 sessions with four coding agents and a tool-free model, none is accepted, and all 35 sessions the simulated user signed off fail at least one program veto.
Université de Montréal Mohamed bin Zayed University of Artificial Intelligence University of Manchester Stevens Institute of Technology
Yupeng Cao（First author）
FinCosmos Agent evaluation Financial Backtesting Look-Ahead Bias
arXiv Oct 8, 2026
Sequential resetting procedures deliver horizon-independent FDR control for LLM watermark detection and financial backtesting
The authors propose sequential resetting procedures built on e-values and test supermartingales that reset the capital to one after each rejection and continue testing, thereby reporting multiple intervals containing null violations; a refinement rule reports only observations after the within-block minimum of the capital. Under independent null e-values the FDR bound is α, while under dependence between null and non-null data the bound is α times a logarithmic factor with anytime validity, demonstrated on LLM watermark detection and ES backtesting.
The authors propose sequential resetting procedures built on e-values and test supermartingales that reset the capital to one after each rejection and continue testing, thereby reporting multiple intervals containing null violations; a refinement rule reports only observations after the within-block minimum of the capital. Under independent null e-values the FDR bound is α, while under dependence between null and non-null data the bound is α times a logarithmic factor with anytime validity, demonstrated on LLM watermark detection and ES backtesting.
University of Waterloo Georgia State University Massachusetts Institute of Technology
Qiuqi Wang（First author）
Sequential Resetting Procedure False Discovery Rate Control E-variable LLM Watermark Detection Financial Backtesting

You’re all caught up