Can You Forecast Crypto Prices With ML? Here Is Competition For You!

Cryptocurrencies have been perhaps the most talked about topic of this decade. Many regard it as a new investment class, some even go as far as comparing it to Gold and calling it Digital Gold. Others regard it as nothing more than a speculative bubble. Either way, the volatility of Cryptocurrencies is something that people on both sides of the aisle accept. G-Research is one of Europe’s leading quantitative finance research firms. G-Research is organizing a competition on Kaggle in partnership with Cambridge Spark asking Machine Learning practitioners to build models to forecast the short-term return of cryptocurrencies. Cambridge Spark is an organization specializing in data science and AI education for workforces.

forecast Cryptocurrencies
Image by Sergei Tokmakov Terms.Law from Pixabay

As a part of the competition, you would be asked to forecast the price of cryptocurrencies in the short term. To know more about the challenge, you can also check out the video below by G-Research-

The Competition

In this competition, you will be provided with a training dataset that will consist of millions of rows of high-frequency market data dating back to 2018. You will have to train a model that will be able to effectively forecast the price of cryptocurrencies.

The results are to be submitted by a python time-series API which will be provided. This will ensure that models do not peek forward in time.

Following the final submission, all the models will be tested on real-world data for 3 months by making the model forecast the price of 14 cryptocurrencies. The winner will be announced after this 3 months of rigorous testing in the real-world.

The 14 cryptocurrencies which are a part of the training data are- Bitcoin, Bitcoin Cash, Binance Coin, EOS.IO, Ethereum, Ethereum Classic, Litecoin, Monero, TRON, Stellar, Cardano, IOTA, Maker, and Elon Musk’s favourite Dogecoin.

The solutions will be evaluated on the weighted version of the Pearson correlation coefficient.

Timelines

The competition began on 2nd November 2021. And while the final submission deadline is on 1st February 2022, the deadline for entry into the competition and team merger is 25th January 2021. So far, there are already 955 competitors from 843 teams who have made 4374 entries.

Post final submission, the submissions will also be tested on data from the real world for a period of 3 month. The competition will end with the announcement of the winner 3 months later on 3rd May 2022.

Prizes

G-Research is handing out big prizes for being able to accurately forecast the price of 14 Cryptocurrencies. The following prizes will be awarded to the top 10 finishers in the competition.

PositionPrize
1st$50,000
2nd$20,000
3rd$15,000
4th$10,000
5th to 10th$5,000 each

This makes the total prize sum being awarded in the competition $125,000.

To be a part of the G-Research Crypto Forecasting challenge, you just need to have a Kaggle Account. To know more about this challenge and participate in it, click here.

Good Luck!


Other Challenges

Check out the Google Tensorflow Challenge for detection and localization of coral-eating crown-of-thorns starfish and help protect the Great Barrier Reef, here.