Weekly Research Recap
Latest research on investing and trading
Welcome to this week’s briefing, your curated roundup of the most actionable investing insights from the past seven days, drawing from academic research, industry reports, blogs, and social media, with links to everything.
Commodities
Latent Characteristic Factors in Commodity Markets (Sakkas and Stavroglou)
Many commodity signals tend to reflect the same underlying forces. Using 11 characteristics across 25 commodity futures, this paper finds that much of their variation can be reduced to just two factors: One related to how risk is shared among market participants, and another related to broader macroeconomic conditions. Key takeaway: The commodity factor zoo seems to be much smaller than it appears.
Crypto
Agentic AI Nowcasting and Cryptocurrency Return Predictability (Sun, Wang, and Zhang)
Agentic AI can turn real-time public information into investable signals. In a real-time test across coins in the CoinMarketCap 100 Index, AI-generated scores predicted crypto returns from 1-hour to 1-week horizons. A Top 10–Bottom 10 portfolio earned 1.78% at the 1-day horizon. Key takeaway: AI can uncover return-relevant information not yet fully reflected in crypto prices.
Equities
Factor Neutralization and Risk-Budgeted Scaling: Evidence from Long-Short Anomaly Portfolios (Kosmakov)
Factor neutralization can improve anomaly portfolios, but only when the risk reduction outweighs the return sacrificed. Across 133 long–short strategies, removing exposure to the Fama-French five factors plus momentum raised average Sharpe from 2.25 to 3.57. Key takeaway: Neutralization is most effective when factor exposure contributes more to portfolio risk than to expected returns.
Fixed Income
Overnight Corporate Bond Trading (Abdi, Rossi, and Wu)
Corporate bond trading contains an unusual overnight signal. When public trade reporting is temporarily paused, institutional trades become much larger, and their order flow carries substantially more information about future prices than daytime trading. The signal even spills across bonds from the same issuer. Key takeaway: Overnight order flow contains information about future bond returns that is only gradually incorporated into prices.
Machine Learning and Large Language Models
International Yield Curves and Exchange Rates (Wang and Zhu)
Exchange rates are more predictable than commonly thought. Using machine learning and yield curves from 17 countries, this paper beats the random-walk benchmark OOS for 8 of 9 G10 currencies at 1 month, and all 9 at longer horizons. Key takeaway: Currency predictability depends on information across global yield curves, not just bilateral interest-rate differences.
Machine learning for realised volatility forecasting (Rahimikia and Poon)
Machine learning can improve volatility forecasts, but its advantage disappears when volatility becomes extreme. Across 23 NASDAQ stocks, ML models using order-book, news, and historical volatility data beat traditional HAR models during 90% of the out-of-sample period, but underperformed on extreme-volatility days. Key takeaway: Combining ML with traditional models is more robust than replacing them.
On the predictability of ETF returns with technical predictors (Gong and Muller)
This paper trains a random forest model on decades of individual-stock data, then transfers the learned relationships to international equity ETFs. The resulting top-minus-bottom decile portfolio earned 0.76% per month out of sample, with momentum and volatility signals particularly effective. Key takeaway: Predictive patterns learned from individual stocks can transfer to ETFs.
Machine learning can improve Treasury yield-curve forecasts, but the gains are highly concentrated. Using 2015–2025 out-of-sample data, the greatest improvements appear at short maturities and in curve-slope forecasts, especially when macro data are included. Key takeaway: Macro information is most valuable for predicting changes in the shape of the yield curve, particularly its slope.
Do Market Regimes Improve Machine-Learning Stock Ranking? (Pascual Miralles and Alfeus)
Machine learning can rank stocks effectively without elaborate regime timing. An XGBoost model ranking 200 U.S. stocks produced a 34.2% CAGR and 1.28 Sharpe out of sample from 2021–2025. Adding market regimes or portfolio optimization failed to improve results. Key takeaway: Adding sophistication to a strong investment signal doesn’t necessarily improve performance.
Prediction Markets
Prediction Markets as Event Hedges: Can They Actually Be Used in Practice? (Kirillov)
Prediction markets can potentially hedge event risk, not just forecast it. A Trump-win contract reduced the one-day loss on a $100k solar ETF position from −9.32% to +0.53%. But limited liquidity made the approach impractical at institutional scale. Key takeaway: Prediction markets can provide targeted event hedges for smaller portfolios.
Blogs
Revisiting Intraday Momentum (Quantseeker)
Bond indices and systematic duration management (Macrosynergy)
Most Asset Managers Already ‘Use AI.’ Few Turn It Into Alpha (CFA Institute)
Podcasts
How a 50-Year Veteran Thinks About Risk Management · Peter Brandt (Chat with Traders)
Trading 80 Synthetic Markets with Trend, Carry & Skew: Jiro Fujisawa, Asset Management One USA (RCM Alternatives)
Portable Alpha: Asking the Questions That Matter ft. Harry Moore (Top Traders Unplugged)
Nicolas Mirjolet, CEO of Quantica on Why the Most Profitable Trends Face the Most Skepticism (StrandGlobalMacro)
Social Media & Industry Research
Traders on Defense (Citadel)
Skewness as a Hidden Driver of Anomaly Returns (Alpha Architect)
Portable Alpha: Ask the Hard Questions (Man Group)
Last Week’s Most Popular Links
Disagreement on Volatility (Yang and Zhu)
A Seat at the Table: Turning AI Macroeconomic Views into Portfolios with a Factor Framework (Elkamhi and Lee)
The Costs and Benefits of Leveraged ETFs (Murray and Sammon)
Disclaimer: This newsletter is for informational and educational purposes only and should not be construed as investment advice. The author does not endorse or recommend any specific securities or investments. While information is gathered from sources believed to be reliable, there is no guarantee of its accuracy, completeness, or correctness.
This content does not constitute personalized financial, legal, or investment advice and may not be suitable for your individual circumstances. Investing carries risks, and past performance does not guarantee future results. The author and affiliates may hold positions in securities discussed, and these holdings may change at any time without prior notification.
The author is not affiliated with, sponsored by, or endorsed by any of the companies, organizations, or entities mentioned in this newsletter. Any references to specific companies or entities are for informational purposes only.
The brief summaries and descriptions of research papers and articles provided in this newsletter should not be considered definitive or comprehensive representations of the original works. Readers are encouraged to refer to the original sources for complete and authoritative information.
This newsletter may contain links to external websites and resources. The inclusion of these links does not imply endorsement of the content, products, services, or views expressed on these third-party sites. The author is not responsible for the accuracy, legality, or content of external sites or for that of any subsequent links. Users access these links at their own risk.
The author assumes no liability for losses or damages arising from the use of this content. By accessing, reading, or using this newsletter, you acknowledge and agree to the terms outlined in this disclaimer.
Paid subscriptions may not be available in all jurisdictions and may change without notice.