Home Community Crypto-ML Roadmap and Feature Suggestions Crypto-ML Upcoming Features [September Update]

This topic contains 2 replies, has 3 voices, and was last updated by Justin Justin 2 weeks, 5 days ago.

  • Author
    Posts
  • #5653
    admin
    admin
    Keymaster

    We are excited to share an update on our upcoming features and development roadmap.

    Machine Learning

    Model series 3.5 [September 6]

    • Trailing stop losses to better capture profits
    • Trade Meter enhancement to sync with various “sell” parameters, including ML sell, predictive sell, stop loss, and trailing stop

    Model series 4.0 [September 20]

    • Integration of trade models with Market Index. Models will consider Market Index trend and value when opening and closing positions. This achieves the initial vision and enhances performance in numerous market conditions, namely during weakening markets.

    Backend

    General

    • The API, database, and web services team has been focused on platform performance and security. This has included refactoring of the entire platform.

    Auto Trade Expansion

    • Binance [October]
    • Bitfinex [TBD]

    Future Development

    • Portfolio management with Auto Trade (multiple currencies)
    • Day trading signals

    Please let us know your feedback and questions below.

  • #5658
    Avatar
    matthew55
    Participant

    Hello, thank you for the update… many interesting features in it!

    I’d like to ask how you’re going to deal with the daily signals, since so far it’s been an area where I’ve almost always lost money 😀

    Are you saying that your algorithm is (or has become) so good that it’s able to predict also much smaller price variations compared to now?

    And will these signals be handled by the Auto Trade plan as well (meaning we won’t have to manually trade every day)?

    Thank you as always, Matteo

    • #5660
      Justin
      Justin
      Participant

      Hello @matthew55, thank you for the response!

      For the day-trading, you’re not alone. 🙂 Doing so is very challenging even for machines.

      What I can say is this is a completely different machine learning model. Our current models really prefer to trade in the “swing trading” time frame–that’s not a constraint we put on the models, but they always settle in that timezone.

      Technically, it’s a big challenge to do the day trading. We’ve had the concepts in place for a while. And our systems have been collecting data specifically for this, but it wasn’t until we put the recent performance enhancements in place that we really felt comfortable moving forward.

      And it’s interesting, in the “swing trading” time frames, our models tend to follow momentum trading behavior. That is, after a trend establishes, they hop in and try to capture a portion of it.

      But for the “day trading” time frames, the opposite seems to be favored. It tends to act on reversals. So this is a completely different tactic–basically looking for smaller bounces to capitalize on.

      To do this right, some of our other recent enhancements (stop losses and trailing stops) will be key.

      Eventually, yes–it would be great to get this in Auto Trade, but it’ll need to be a stair-step approach.

      I envision this as you want to sit down and trade for the next 3 hours. You can pull up our system and watch it as it feeds you data and levels. And hopefully you walk away with some solid profit!

You must be logged in to reply to this topic.