Poll: Next Machine Learning Project
We have just wrapped a major upgrade of our signals to our next-gen machine learning models. Before we officially kick off our next task, we’d love to get your input!
Please note that this poll is specific to the machine learning work and does not include roadmap work done by our web team and API team. Those teams will continue to work separate enhancements such as revising the Member Dashboard, opening API access, and developing Portfolio Management functions for Auto Trade.
Additionally, while our machine learning team’s top priority is to monitor and improve our core signals, we also want to continuously work on what’s next.
With that said, please take this opportunity to help drive our roadmap!
Next machine learning feature poll
The survey has ended! See the results here:
Here is additional detail on the features listed above:
Bitcoin Dominance Predictor: A machine learning model would offer a prediction on how Bitcoin will perform relative to altcoins. Historically, this has gone in waves. Bitcoin has had strong runs followed by altcoins (aka “Alt Season”). Having a prediction here would help you adjust your portfolio and better capture these waves.
Short selling: We have previously provided a trade signal optimized for short selling Bitcoin. This is currently the only model not upgraded to the 4.x architecture. We can upgrade this, but it is extensive work and unique from the other models. Therefore, before prioritizing this just because we’ve done it in the past, we want your opinion.
Maximize BTC value: For machine learning models to be effective in trading, they need to have one clear goal. For our models, that goal is to maximize the account value in USD terms. This is why our Trade History page is expressed in USD terms. While this may result in higher BTC holdings over time, it does not necessarily. This option would be for those looking to maximize their held BTC quantity instead of their USD value.
Crypto-to-crypto pairs: Rather than trading against USD (or USD stablecoins), this option would be to trade BTC against other cryptocurrencies. We would examine the most popular pairs and look to generate models accordingly. This would ultimately pair well with the Bitcoin Dominance Meter model.
Low-frequency signal: Our currrent signals may issue several signals per month, or even several signals per week. We could pursue a model that has a longer time horizon and is more ideally suited for investors (rather than traders).
Do you have a different suggestion?
Please let us know in the comments below or reply on our Community Forums Roadmap post.
Crypto-ML provides machine learning for crypto traders and investors. Gain crystal-clear signals and deep market insights. Add predictive capabilities to your toolbox. Learn more and join for free.