Numerai Tournament Achieves Market Efficiency Milestone

馃幆 When AI Beats Humans

By Numerai
Nov 28, 2024, 5:15 PM
twitter
News article
Photo by Numerai

The Numerai Tournament has reached a significant efficiency milestone with its Stake-Weighted Meta Model.​ Key highlights:

  • 413 data scientists actively contributing
  • $7 million total staked
  • Meta Model outperforms 99% of individual contributors

This follows recent growth in Numerai Crypto division:

  • Over $2M staked across 312 models
  • Free predictions available for top 500 tokens
  • Meta Model shows highest correlation with ranked returns

View performance data and predictions at crypto.​numer.​ai/meta-model

Sources

Amazing work by Numerai participant, jefferythewind. "Following up on some exploratory work on simulating portfolios built from Numerai鈥檚 crypto meta model, I wanted to investigate a statistical risk factor model for the crypto market and apply the model to more realistic

47
Reply

The Numerai Tournament is itself an efficient market. The Stake-Weighted Meta Model is powered by 413 data scientists staking $7 million, and it is currently outperforming 99% of individual contributors.

Richard Craib
Richard Craib
@richardcraib

numer.ai/~numerai_swmm meta model beats all

53
Reply
Read more about Numerai

Numerai Launches Crypto V2.0 Spectra Dataset for Prediction Markets

Numerai Launches Crypto V2.0 Spectra Dataset for Prediction Markets

**Numerai has released its Crypto V2.0 Spectra Dataset**, expanding the platform's prediction capabilities into cryptocurrency markets. The new dataset follows the recent launch of Numerai Signals V2 "Cosmic" dataset, which featured: - Massively expanded universe of assets - Improved ticker consistency - Enhanced data quality for predictions **Key details:** - Available immediately for data scientists and quants - Builds on Numerai's prediction tournament model - Users can stake NMR tokens on their crypto predictions This release represents Numerai's continued expansion beyond traditional equity markets into digital assets, providing structured data for machine learning models to predict cryptocurrency performance. The platform allows participants to submit predictions and stake NMR tokens based on their confidence levels, creating a decentralized hedge fund powered by crowd-sourced intelligence. [Read the full details on Numerai's forum](https://forum.numer.ai/t/crypto-v2-spectra-dataset/8197)

NumerCon 2026 Scheduled for January 30th in San Francisco

NumerCon 2026 Scheduled for January 30th in San Francisco

**NumerCon 2026** is set for **January 30th** in San Francisco. The annual conference brings together the Numerai community of data scientists and machine learning practitioners. - Date: January 30th, 2026 - Location: San Francisco - Event: NumerCon 2026 Mark your calendars for this key gathering in the prediction market space.

Decentralized AI Day Vienna

**Decentralized AI Day** is coming to Vienna on **September 20th**. The event will feature: - Discussions on **Numerai** developments - Latest trends in **decentralized AI** - Networking with industry professionals Hosted by @NumeraiCoE, this gathering follows successful events in other cities. **RSVP required**: [Register here](https://luma.com/5w9wvbjf?tk=8Bi8Mv) Don't miss this opportunity to connect with the decentralized AI community.

Numerai Signals Switches to Alpha + MPC Scoring

**Numerai Signals is implementing a major scoring change** starting September 2nd. The platform will transition from its current system to **Alpha + MPC scores** for all new rounds. - Alpha scoring was introduced in August as a new evaluation method - MPC (Multi-Party Computation) scores will be combined with Alpha - Change affects all prediction rounds beginning September 2nd or later This represents a significant shift in how Signals evaluates and rewards predictions on the platform.

Numerai's Open Data Strategy: Free Hedge Fund Training Data for AI Models

**Numerai's Revolutionary Approach** Numerai hedge fund operates on an unusual model: **giving away all trading data for free** to anyone who wants to train machine learning algorithms on it. The platform allows data scientists to: - Access proprietary financial datasets at no cost - Train their own ML models using this data - Submit predictions directly to the hedge fund - Potentially earn rewards for accurate forecasts **The Bigger Vision** Numerai's 2017 master plan positioned the platform as an **API for AI capital control** - essentially becoming infrastructure that any artificial intelligence could use to manage money in the economy. This open approach means Numerai stays **permanently accessible to new ML techniques** and can absorb the best innovations from the global data science community. **Why This Matters** While most hedge funds guard their data jealously, Numerai's model flips traditional finance on its head by crowdsourcing intelligence rather than hoarding it. Read more about their strategy: [Numerai's Master Plan](https://blog.numer.ai/numerais-master-plan/)

other