Automate participation in the Numerai hedge fund tournament to train models and earn NMR cryptocurrency.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install numerai-tournament
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install numerai-tournament using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Numerai Tournament?
The Numerai Tournament skill enables AI agents to participate autonomously in a global data science competition hosted by a quantitative hedge fund. Users can leverage this skill to download obfuscated financial datasets, train sophisticated machine learning models using LightGBM, and submit predictions to earn NMR cryptocurrency based on market performance.
By integrating this functionality into Openclaw Skills, developers can bridge the gap between raw data science and automated financial execution. The skill handles the complexities of the Numerai API, data versioning (v5.2), and the specific runtime requirements needed for successful tournament participation, allowing for a hands-off approach to quantitative trading strategies.
Numerai Tournament Use Cases
- Automating weekly stock market prediction submissions to the Numerai hedge fund.
- Building an autonomous pipeline for machine learning model training on obfuscated financial data.
- Developing a passive income stream through NMR cryptocurrency staking and performance rewards.
- Testing ensemble learning strategies across multiple financial targets and feature sets.
How Numerai Tournament Works
- The skill authenticates with the Numerai API using provided public and secret keys stored in the environment or a local credentials file.
- It synchronizes the latest v5.2 tournament data, including training, validation, and live parquet files from Numerai servers.
- A LightGBM ensemble is trained on specified targets (e.g., target_teager2b_20) using a selected feature set size (small, medium, or all).
- Predictions are generated for the live dataset and rank-normalized to ensure compatibility with the Numerai meta-model.
- The skill either uploads a CSV of predictions manually or deploys a pickled model function for zero-maintenance automated daily execution.
Numerai Tournament Setup
Install the necessary Python dependencies and configure your environment variables to start using this within Openclaw Skills:
pip install numerapi lightgbm pandas numpy cloudpickle scikit-learn
export NUMERAI_PUBLIC_ID="YOUR_PUBLIC_ID"
export NUMERAI_SECRET_KEY="YOUR_SECRET_KEY"
For macOS users on Apple Silicon, ensure the OpenMP library is available for LightGBM:
brew install libomp
Numerai Tournament Data Schema & Taxonomy
The skill manages data within a local directory structure, primarily utilizing Parquet format for performance and memory efficiency:
| File | Description |
|---|---|
train.parquet |
Historical data for model training (~8GB) |
validation.parquet |
Data for backtesting and Sharpe ratio calculation |
live.parquet |
Real-time features for the current tournament round |
features.json |
Metadata defining feature sets (small, medium, all) |
ensemble_models.pkl |
Pickled dictionary containing trained target-specific models |
Numerai Tournament Advanced Features
- Automated Model Uploads: Deploy a pickled function directly to Numerai infrastructure for 24/7 autonomous prediction without needing a local server.
- Multi-Target Ensembling: Simultaneously train on diverse targets like target_cyrusd_20 to improve ensemble diversity and overall Sharpe ratio.
- Feature Neutralization: Advanced logic to decorrelate predictions from common market factors, boosting Meta-Model Contribution (MMC) within Openclaw Skills.
- Era-Aware Validation: Integrated performance tracking that evaluates model correlation per-era to ensure stability across non-stationary financial data.
SKILL.md
Loading
Related Openclaw Skills
Big Memory
An advanced context management system that uses structured snapshots to ensure AI agents never lose track of complex coding tasks during context compaction.
Sendook Email Skill
A specialized skill for AI agents to interact with Sendook inboxes by reading, searching, replying to, and sending emails.
Sendook Email Skill for Openclaw
A specialized skill for AI agents to interact with existing Sendook email inboxes for automated messaging and thread management.
GTA Property Investment Report Generator
An automated tool for generating professional property investment financial reports specifically for the Greater Toronto Area.
ObekT Security
A comprehensive threat detection and security analysis toolkit designed to audit code, files, and AI agent skills for vulnerabilities and malicious patterns.
WOL and Sleep PC Control
A specialized skill for sending Wake-on-LAN and Sleep-on-LAN magic packets to manage PC power states across a local network.
Featured*
