Relevance
5/10
Directly addresses portfolio optimization (Markowitz framework) with cardinality constraints, a core quantitative trading problem. However, the paper explicitly does not claim quantum advantage and shows the penalty-free pipeline is competitive with but not superior to greedy classical references. The primary contribution is a structural diagnosis (penalty encoding, not hardware, is the bottleneck) rather than a practical trading edge. The betting case study and block-diagonal structures have limited direct trading application. The finding that classical post-processing dominates the pipeline quality is important context for practitioners considering quantum approaches.
Implementation Complexity
6/10
The penalty-free pipeline itself is conceptually simple (remove penalty, sample objective-only QUBO, apply greedy projector). However, full replication requires: D-Wave Leap access (Advantage_system4.1 Pegasus and Advantage2_system1.13 Zephyr), understanding of minor embedding (minorminer, FixedEmbeddingComposite), QUBO formulation, chain-strength sweeps, and the greedy feasibility projector. The sparsification analysis (4 families) and ablation studies add complexity. The offline experimental grid (918 preservation rows, 4,212 embedding rows) is computationally intensive. Code is available but requires quantum computing infrastructure access.
Reproducibility
4/5
Code released under MIT License at https://github.com/LuisLozanoM/penalty-free-portfolio. Uses publicly available data (Fama-French 49-industry portfolios, football-data.co.uk). Fixed random seeds reported. Live QPU experiments on D-Wave Leap platform (Advantage_system4.1 and Advantage2_system1.13). However, live QPU results depend on hardware state and may not be exactly reproducible. Detailed parameter settings provided in Table 1. Extensive supplementary materials (Online Resources A-G) included.