Evaluate predictive out-of-sample trading performance

Evaluate the predictive trading performance and practical viability of the tensor-network field-coupled XY portfolio construction through an out-of-sample analysis, including the effects of parameter selection and trading costs.

Background

The empirical comparisons in the paper use the same estimation sample for the XY portfolios and benchmark strategies. The study therefore establishes computational feasibility and describes in-sample risk–return behavior, but it does not test whether the inferred scores and softmax allocations generalize to future observations.

The paper also identifies omitted trading costs and the use of same-sample parameter inspection as limitations. A concrete unresolved question is whether the method produces economically meaningful predictive or trading performance under rolling or otherwise properly out-of-sample evaluation.

References

Predictive performance, statistical superiority, and trading viability require an out-of-sample analysis.

Tensor-Network Inference in a Field-Coupled XY Model for Portfolio Allocation  (2609.05045 - Chowdhry et al., 4 Sep 2026) in Section 1, Introduction; Section 6, subsection “Empirical interpretation and scope”; Section 7, Conclusion