Numerai is a data science tournament. It’s like Algorithmic trading without doing the trading part.
Tournaments
Numerai has two tournaments.
Classic / Regular
- Try to predict on a black-box dataset.
-
You don’t know the stock tickers or what the data features mean.
- ID:
1a868e3e-1df7-45ca-91e0-34e315f5bee2
- Name: kazutsugi
- Tournament: 8
Signals
- You bring your own data and predict for each stock ticker.
-
You know which stocks you are predicting for.
- ID:
1adb7bf1-f61a-4718-a73b-fd516e32a475
- Name: signals
- Tournament: 11
Crypto
- This is the newest Numerai tournament.
- You bring your own data and try to predict Cryptocurrency returns.
Classic
In this tournament, you get an obfuscated dataset. You don’t know the stock tickers, and all the features are normalized into 5 distinct values and have random names.
Datasets
v4.3
That’s a lotta features
v5
The v5 dataset has the same features as v4, but with an expanded universe.
Signals
API
Numerai has a GraphQL API. Both public data (like rounds, dataset URLs, leaderboards etc) and private data (authentication, submissions etc) go over a single endpoint.
The GraphQL endpoint is https://api-tournament.numer.ai
.
To play around with the API, you can use a dedicated GUI like GraphQL Playground. You can also simply visit the endpoint URL with a browser, there is a built-in GUI you can use.
NMR (Numeraire)
NMR is the cryptocurrency / utility token used in the Numerai tournaments.
Website
Name | Link |
---|---|
Main website | https://numer.ai |
Documentation | https://docs.numer.ai |
Forum | https://forum.numer.ai |
Users
Numerai user | Discord user | Notes |
---|---|---|
0xleo | Leo | |
Zyberal | Joo |
Links
- https://github.com/numerai/example-scripts
- https://colab.research.google.com/github/numerai/example-scripts/blob/master/feature_neutralization.ipynb