Quant

AlphaAgent tackled “LLM-generated factors look too similar, accelerating alpha decay” with factor-level regularization. QuantaAlpha (arXiv 2602.07085) operates one level up: instead of constraining a single generation, treat the full “hypothesis → factor → backtest” pipeline as a trajectory, and apply mutation and crossover between trajectories. On CSI 300 with GPT-5.2 it lands IC 0.0472, ARR 4.68%, MDD 11.8%. Zero-shot transfer to CSI 500 and S&P 500 keeps 40.28% and 19.1% cumulative excess return respectively. ...

0DTE options (Zero Days to Expiration) are option contracts that expire on the same day they trade. In 2022, CBOE filled out the SPX expiration calendar so that contracts expire on every weekday (Tuesdays added in April, Thursdays in May), moving 0DTE from a niche product into mainstream trading. Cboe public materials report that 0DTE contracts account for more than half of daily SPX option volume, with the share continuing to rise through 2024-2025. Public discussion of 0DTE tends to cluster at two extremes: social-media narratives of rapid wealth, and warning labels of indiscriminate risk. This article takes neither stance. The aim is to examine 0DTE as a financial instrument: its pricing mechanics, the evolution of Greeks during the final trading day, the payoff structures faced by the main participant types, and the empirical findings of existing academic work on the realised return distribution of retail buyers. With these in hand, a reader can form an independent view of whether the instrument fits their objectives. ...

Qlib is Microsoft’s open-source AI quant platform, released in 2020 and now sitting at 16k+ stars on GitHub. It bundles data, factors, models, backtesting, and reporting into one pipeline, ships with the Alpha158 and Alpha360 factor sets, and includes 20+ models out of the box (LightGBM, Transformer, TCN, and more). This Qlib tutorial walks from installation to a working LightGBM + Alpha158 strategy, then to reading the backtest report, and ends with a checklist of pitfalls. By the end you will: have Qlib running locally, have downloaded A-share daily data, train a model in 30 lines, run the standardised qrun workflow, and understand IC, annualised return, and max drawdown numbers in the output. ...

This vectorbt tutorial starts from the real difference between vectorized and event-driven backtesting, then walks through the three Portfolio constructors (from_holding, from_signals, from_orders), runs a 10,000-combination grid search, and ends with the freq, signal-shift, and memory traps that quietly produce wrong numbers. ...

The previous AlphaGPT review left an open question: when everyone uses LLMs to mine factors, how long can those factors stay effective? AlphaAgent (paper, KDD 2025) tackles this head-on. Its core observation: LLM-generated factors lean too heavily on existing knowledge, producing homogeneous signals that crowd the same trades and accelerate alpha decay. The fix is three regularization constraints injected into the factor generation process, forcing the model to explore structurally novel, logically coherent, and complexity-controlled factors. ...

Quant trading articles are full of math terms: log returns, standard deviation, covariance matrices, OLS regression, partial derivatives. You can look up each definition, but without the trading context, the definitions don’t stick. This article goes through the most common math concepts in quant trading, each anchored to a real trading scenario with Python code. ...

A stock returned 20% last year. How much of that came from the overall market rising, how much from it being a small-cap, and how much from it being cheap? The Fama-French factor model is the tool that answers this question. It decomposes stock returns into a handful of explainable “factors” and forms a foundational framework for quantitative stock selection. This article starts from CAPM, builds up to the three-factor and five-factor models, and runs a factor regression in Python. ...

Quantitative trading is not just about building strategies. Picking the wrong market kills more quant projects than picking the wrong signal. A mean reversion strategy that works on US equities might blow up on crypto because the volatility is 3x higher. A trend following system profitable on commodity futures might sit idle for a year on EUR/USD because major currency pairs barely trend. The market itself determines what strategies survive: its rules, data quality, leverage structure, and trading hours all matter. This article breaks down the five major markets from a quant trader’s perspective. ...

CTA (Commodity Trading Advisor) is a broad category of quantitative strategies that trade futures. Despite the name suggesting commodities only, CTA strategies cover equity index futures, treasury futures, and FX futures as well. The space splits into two camps: trend following (go long when prices rise, short when they fall) and statistical arbitrage (find mispricings between related instruments and bet on convergence). This article covers both, with Python code for each. ...

The core logic of pairs trading is straightforward: find two stocks whose prices “should move together,” and when they diverge from their historical relationship, bet on convergence. Go long the relatively cheap one, short the relatively expensive one, close when the spread reverts. You are not betting on market direction. You are betting on the relative pricing between two assets returning to normal. ...