This project addresses the challenge of extracting actionable trading intelligence from unstructured crypto influencer content. It distills the experience of 99 top crypto traders into 87 backtestable quantitative factors using LLMs, offering quantitative traders and researchers a systematic way to leverage expert intuition. The core benefit is access to data-driven trading signals derived from collective wisdom.

How It Works

LLMs process 39,843 tweets and 17,657 K-line screenshots from 99 selected crypto traders to distill trading insights. Approximately 470 "trading intuitions" are extracted and transformed into 87 Python-based quantitative factors. These factors are backtested against 832 days of historical price data for cryptocurrencies and macro indicators, identifying statistically significant signals and quantifying individual trader reliability.

Quick Start & Requirements

• Primary run command: python3 quant_factors/run_consensus.py --refresh-ohlc or via Claude Code skill /consensus.
• Prerequisites: Python 3.x, historical price data (ohlc_daily.json), and macro data (macro_daily.json) as structured within the repository.
• Links: No direct links to official documentation, demos, or community forums are provided in the README.

Highlighted Details

• Distills insights from 99 crypto KOLs into 87 backtestable quant factors, including high-IC signals like "200W MA Value Zone" (+0.297 IC) and counter-signals like "Gold Hedge" (-0.218 IC).
• Quantifies trader reliability, classifying 58 traders as "followable" and 41 as "counter-signal" based on stated methods vs. historical performance.
• Provides real-time consensus output (bullish/bearish, vote counts, price ranges) with backtested performance showing a 57% win rate on BTC daily signals.
• Features extensibility for adding traders via LLM distillation, integrating additional data sources, and incorporating custom Python factors.

Maintenance & Community

The README provides no specific details on contributors, sponsorships, community channels (e.g., Discord, Slack), or a public roadmap.

Licensing & Compatibility

Licensed under the MIT license, which is permissive for commercial use and closed-source linking.

Limitations & Caveats

This is explicitly a research project, not trading advice. Backtest results do not guarantee live performance, and the 7-day validation sample lacks statistical significance. Users assume all risks.