Blog
Insights on quantitative investing, market analysis, and systematic trading strategies.
Loading articles...
The 60/40 Portfolio Has a Regime Problem
A 2026 review article makes the case that static 60/40 allocation is structurally exposed to regime changes. The fix is not to abandon diversification but to make it conditional on the macro state.
Read moreWhen Alpha Breaks: Why Your Ranking Model Needs a Kill Switch
A new paper by Ursina Sanderink shows that cross-sectional stock rankers can fail silently under regime shifts, and that standard uncertainty measures like VIX are useless for detecting it. The proposed fix is surprisingly simple: a binary trade-or-not gate built from the model's own trailing performance, combined with a discrete tail-risk cap for the most uncertain predictions.
Read moreAn AI That Discovers Its Own Trading Factors. Does It Actually Work?
A recent paper from HKUST proposes an autonomous AI system that generates, tests, and refines its own equity factors without human guidance. The reported Sharpe ratio of 3.11 is exceptional, but the details raise questions any systematic investor should think through before getting excited.
Read moreWhat Quant Traders Actually Need to Get Right: Trends and Risk
A recent survey paper consolidates the evidence on market trend characteristics and risk management for quantitative strategies. The practical conclusion is straightforward: adaptive trend identification matters, but multi-layered risk controls are what keep you in the game.
Read moreA P/E of 200 Beat a P/E of 20. Here Is Why That Should Not Surprise You.
Everyone learns that low P/E means cheap and high P/E means expensive. A recent paper walks through why that logic fails, using Broadcom and Applied Materials as a case study. The answer involves a metric from 1984 that most investors have never heard of.
Read moreMomentum Crashes Are Optional (If You Let a Machine Drive)
Momentum strategies print money until they don't. A new paper shows that a simple LightGBM model can keep the upside and cut max drawdown from 30% to 13%. The trick is knowing when to stop listening to price and start listening to fundamentals.
Read moreWhy the Crowd Gets It Wrong: Herding Behavior and What It Means for Portfolio Construction
A recent study confirms what systematic investors have long suspected: most individual investors follow the crowd, and it costs them. The data on herding, psychological biases, and their impact on diversification.
Read moreTrading Volume Alpha: Why Predicting Volume Matters More Than You Think
A recent paper by researchers at Yale, McGill, Johns Hopkins, and HKUST demonstrates that predicting trading volume with machine learning can be as valuable as predicting returns themselves, doubling Sharpe ratios for billion-dollar portfolios.
Read moreNo articles published yet. Check back soon.