$ show xingwudao/xquant-beginner
This repository contains the open-source book 'XQuant: δΊΊδΊΊι½ζ―ιεδΊ€ζε' (XQuant: Everyone Can Be a Quantitative Trader), which serves as an educational resource for learning quantitative trading methodologies. Rather than being a functional trading application, it provides a structured curriculum teaching readers how to develop, test, execute, and monitor quantitative strategies using AI coding assistants. The book walks through a complete workflow: from selecting investment instruments (ETFs), determining position sizing, creating trading rules, validating strategy performance, avoiding overfitting, modeling execution costs, to monitoring and iterating on live strategies. Its unique value lies in teaching a spec-driven development approach where users write clear strategy specifications that AI coding assistants can translate into executable code, making quantitative research accessible to non-programmers.
The LLM identified 4 files worth highlighting for reuse:
Teaches a structured 4-part specification framework (Context, Task Description, Task Requirements, Acceptance Criteria) for communicating quantitative strategy requirements to AI coding assistants. This reproducible methodology ensures clear communication and reduces implementation errors when translating trading ideas into executable code.
Provides a complete framework for evaluating predictive factors using Information Coefficient (IC) analysis, including time-series IC, grouped IC analysis, and ICIR calculation. This teaches researchers how to properly assess factor predictive power and stability beyond simple backtesting.
Details essential overfitting prevention techniques including walk-forward analysis (rolling and anchored approaches), time-series cross-validation, and rule burden analysis. These methods help distinguish between genuine strategy edge and historical overfitting in quantitative research.
Teaches realistic trade execution modeling by layering transaction costs (commission, slippage, stamp duty) and simulating different fill price models. This provides a framework for understanding the impact of execution costs on strategy performance and creating more realistic backtests.