This is a summary of links recently featured on Quantocracy as of Sunday, 07/20/2025. To see our most recent links, visit the Quant Mashup. Read on readers!
-
Carlson’s “Defense First” [Allocate Smartly]This is a test of Thomas Carlsons Defense First strategy from his paper Defense First: A Multi-Asset Tactical Model for Adaptive Downside Protection. Strategy results from 1971 follow. Results are net of transaction costs see backtest assumptions. Learn about what we do and follow 90+ asset allocation strategies like this one in near real-time. Logarithmically-scaled. Click for
-
When your strategy works, is it just dumb luck? How to stack the odds in your favour [Robot Wealth]Recently, we had an excellent question on the Trade Like a Quant Discord server: How do you know if your strategy is working out of coincidence rather than actual edge? The strategy might work over a long period just because of blind luck. Damn good question. It hits right in the insecurity because the honest answer is: you can never know for sure. I remember when I first started trading. I
-
The Memorization Problem: Can We Trust LLMs Forecasts? [Quantpedia]Everyone is excited about the potential of large language models (LLMs) to assist with forecasting, research, and countless day-to-day tasks. However, as their use expands into sensitive areas like financial prediction, serious concerns are emergingparticularly around memory leaks. In the recent paper The Memorization Problem: Can We Trust LLMs Economic Forecasts?, the authors
-
Behavioral Biases and Retail Options Trading [Relative Value Arbitrage]Why Do Investors Lose Money? Behavioral finance is the study of how financial behavior affects economic decisions and market outcomes, and how those decisions and outcomes are affected by psychological, social, and cultural factors. Behavioral finance research has shown that people do not always make rational decisions when it comes to money. Factors such as emotion, social pressure, and cognitive
-
Do Smart Machines Make Smarter Trades? [Alpha Architect]Can machine learning models help us exploit stock market anomalies more effectively? This paper says yesbut with a few important caveats. By applying gradient boosting algorithms to a wide array of established anomalies (like value, momentum, and quality), the authors show that machine learning methods can significantly improve the performance of long-short strategies. These models capture