Rating
1696
Battle Count: 56
Relevance
9/10
Highly relevant to quantitative trading, particularly FX market making and algorithmic execution on aggregator-routed RFQ platforms. The paper directly addresses optimal quoting strategies under inventory risk, adverse selection, and platform-level routing feedback. The campaign-vs-harvest dynamic and hysteresis findings have practical implications for dealer pricing policies. The two-tier structure (gated vs. background flow) mirrors real FX workflows. The HJB-based framework provides a principled approach to multi-size ladder quoting. However, it is primarily a theoretical/analytical contribution rather than a directly deployable trading system.
Implementation Complexity
8/10
High complexity: requires solving multi-dimensional HJB PDEs with non-local jump operators indexed by trade size, implementing adiabatic two-scale decomposition, quadratic inventory ansatz with Riccati closure, logistic gate parameterization, fold bifurcation detection, and fixed-point iteration with Anderson acceleration for stationary solutions. The numerical scheme involves careful handling of timescale separation (inventory relaxation vs. score dynamics), grid resolution for 10^6+ time steps, and convergence of the continuation-value fixed-point iteration. Analytical derivations (envelope theorem, Hamiltonian expansions) require strong mathematical background in stochastic control and PDE theory.
Reproducibility
3/5
The paper provides detailed parameter settings (size ladder, intensities, win-probability sigmoid parameters, gate parameters, EMA alpha, volatility, risk aversion) and describes the numerical scheme (block-interval HJB with fixed-point iteration, Anderson acceleration, T=0.05 days, 10^4 time steps). However, no code repository or implementation is provided. The analytical derivations are self-contained but the numerical PDE solver details (grid, boundary conditions, discretization) are only partially specified. Reproducing the full numerical results would require significant effort.
About this paper
Methodology: Two-Tier Stochastic Optimal Control with Adiabatic Quadratic Approximation. Problem types: Market Making, Optimization, Risk Management, Algorithmic Execution, Stochastic Control.
The interactive Everscope explorer (charts, battles, favorites) loads below.