Research update — 1 review, 1 strategy, 33 new papers

New public strategies

• Regime-Adaptive Student-t GARCH Volatility Targeting for the NEPSE Index

  The strategy holds a long NEPSE Index exposure scaled to a 12% annualized volatility target using one-day Student-t GARCH or EGARCH forecasts. It shortens the estimation window and halves exposure when point-in-time stability diagnostics fl…

  Sharpe 0.83 · total return 37.3% · max drawdown -11.4%

Quant Paper Radar · 10 of 33

1. 1

   What survives honest evaluation? Leakage-safe, search-aware assessment of LLM-driven trading strategy discovery

   Eray Gençay · ARXIV

   we backtested it — Sharpe -0.10 · quant relevance 0.98

   The paper presents a practical framework for evaluating LLM-discovered trading strategies without look-ahead leakage and while accounting for the full number of searched candidates. It reports real-data backtests on US stocks and ETFs, finding that passive benchmarks pass its tests while all evaluated LLM-discovered strategies fail after honest out-of-sample and multiple-testing-aware evaluation.

   original paper

2. 2

   Harvesting the Variance Risk Premium in Nuclear and Energy Equities: A Short-Put Portfolio Derisking Strategy

   Jilang Miao, Nonna Sorokina · ARXIV

   quant relevance 1.00

   The paper tests a systematic cash-secured short-put strategy intended to harvest the single-stock equity variance risk premium in nuclear and energy-adjacent equities. It reports gross backtest performance for unconditional and IV-to-GARCH-realized-volatility-filtered option-writing portfolios.

   original paper

3. 3

   Deep Hedging Under Realistic Market Frictions: A Regime-Conditional Empirical Study of Dynamic Option Hedging on Bitcoin Options

   Sheryan Kumar · ARXIV

   quant relevance 0.99

   This is a real-data empirical comparison of neural deep-hedging policies with classical transaction-cost-aware option hedges. It reports that a Whalley-Wilmott no-trade band materially reduced transaction costs and that the tested deep-hedging models did not outperform the classical benchmarks in its BTC-options test window. The platform cannot reproduce the paper's hourly Deribit BTC-options experiment, but can test the same discrete, cost-aware hedging mechanism at EOD on US listed equity options.

   original paper

The rest of the batch

• End-to-End Neural Shrinkage of Indefinite Pairwise Correlation Matrices for Small-Cap-Inclusive Portfolios

  Christian Bongiorno, Lorenzo Villassero · ARXIV · relevance 0.99

  The paper proposes a neural, rotation-invariant covariance shrinkage method that converts indefinite pairwise-complete correlation matrices into positive-definite covariance estimates for long-only minimum-variance equity portfolios.

• Agentic Quantitative Trading: A Survey of Workflows, Systems, and Evaluation

  Fengrui Hua, Hengyi Yang, Xinlei Hao et al. · ARXIV · relevance 0.98

  This is a survey of LLM-based agentic systems for quantitative trading, spanning factor mining, signal discovery, portfolio construction, execution, and risk management.

• Foundation models for electricity price forecasting and battery arbitrage: Can they replace market-specific forecasting models?

  Arkadiusz Lipiecki, Rafał Weron · ARXIV · relevance 0.96

  The paper evaluates zero-shot foundation models for probabilistic day-ahead electricity-price forecasting and converts forecasts into battery-storage arbitrage decisions in German, Polish, and Spanish power markets.

• Portfolio Risk Bounds without Cross-Asset Return Covariances: Distributional Fields from Language-Model Representations

  Marcus Gawronsky, Chun-Sung Huang · ARXIV · relevance 0.96

  The paper proposes a long-only portfolio-allocation rule that minimizes a Wasserstein-distance-based upper bound on systematic portfolio variance, using distributions of firm-level news embeddings rather than estimated cross-asset return…

• Agentic Empirical Asset Pricing: Methodological Foundations

  Yingjian Pan, Xiaowei Ding, Kay Giesecke · ARXIV · relevance 0.94

  The paper develops and evaluates an LLM-agent system for discovering cross-sectional US equity factors, with point-in-time validation and rolling re-execution of the discovery process.

• The Convergence Rate of Stochastic Tracking with Application to Optimal Execution

  Marcel Nutz, Moritz Voss · ARXIV · relevance 0.93

  This is a theoretical optimal-execution paper for an Obizhaeva-Wang transient-price-impact model with stochastic terminal inventory targets.

• RetailAgent: Structured Adverse Timing in Self-Conditioned Multimodal LLM Trading Agents

  Yupeng Zhang, Liuyuan Jiang, Hongyi Huang et al. · ARXIV · relevance 0.91

  The paper evaluates sequential LLM long/flat equity decisions on intraday price paths and finds persistently adverse within-day timing.

and 23 more · see the full radar