Making Profit on Forex (with proofs) using my Trading Algorithm
In this brief article I will continue what I indicated in my first project published on Medium.
I started creating a neural network that predicted different prices and directions for crypto assets. I have used the same technology (convolutional neural networks) to make price predictions for 20 different Forex pairs.
The results have been so far incredible, I was not expecting getting the levels of accuracy I achieved.
But… What is the difference between crypto & forex in regard to trading hours?
The main difference between forex and crypto trading is that cryptos are available for trading 24x7, while forex market closes during the weekends.
When it opens on Sunday evening, there are normally no price gaps (they are way more common on the stock market), and that was something that really helped my predictions to be accurate.
The algorithm I developed creates every single hour a new price prediction for all of the 20 forex pairs for 5 different timeframes: 1 hour, 4 hours, 1 day, 3 days and 1 week.
Some nice results
It already takes into account the time-gap of weekends, so during the last hours of a Friday, it will start making predictions for Monday. Here I include some weekly yields for the EUR/AUD pair:
You can check them yourself at SlothTrading, where I publish all the data and can be accessed for free. What I have observed is that if I take the yield results for the 1-week timeframe and already consider transaction fees and slippage for an exchange like FXCM, it is profitable in the medium-long term.
That is SURPRISINGLY GOOD!
I invite you to take a look at the website. In a future, my idea is to offer people the possibility of linking their exchanges via API with the data from my algorithm so you can somehow automate the trades, what do you think?
I will keep working to improve the accuracy rate for all pairs, as well as for including predictions for the US stock market (if you have any suggestions send me a mail). See you in the next blogs!
My website, where all the data is published.
Follow me on Twitter.
Give us a Donation to continue developing this project, it would be highly appreciated:).