When I first starting thinking about what type of python based cryptocurrency project I wanted to work on I tried a few simple ideas and quickly became dissatisfied. My early failures and misfires proved to be invaluable however as it led me to what I considered to be the most pertinent question.
What should the price of any given cryptocurrency be at any given point in time?
What is ETH?
Ether is an alternative to Bitcoin. I could have chosen any cryptocurrency to work with but I believe in the Ethereum platform. Indeed my algorithm will work very much the same way with Bitcoin or any other token that has readily available data with only slight modifications to the code.
If you are interested in learning more about ETH and Ethereum in general, this is a great place to start.
What the Price of ETH Should Be
While this title may sound somewhat pretentious; it really is at the heart of what all investors do.
That is, generally, one wants to buy when the price is low and sell when it’s high. If one can determine what the price of anything should be, then it is trivial to figure out if it’s too high or too low.
In the chart above the blue line is what my algorithm determined the price of ETH should be. The orange line is what the actual price of ETH was over time. As you can see my line and the actual price cross each other regularly and there are also points where they diverge sharply.
By identifying the peaks and valleys when the actual price is farthest away from the model of where the price should be, it is simple to determine when you might want to buy or sell.
While this chart shows past performance ( through yesterday ) my system uses a deep learning model called a Recurrent Neural Network (RNN) to do time series prediction. If your model is good enough, and all things being equal, time series prediction is very much like being able to see a few steps into the future.
The State of this Project
I am continuing to train my model and have even begun to “cross train” it (as it were…) on other crypto coins. There is at least some correlation of performance across the entire market.
While my model is continuing to improve, it’s documented performance up until now has been effective. In fact, if someone had been using it religiously to play the ETH market using Fibonacci retracement levels, the golden ratio, etc. one could have made a tidy profit.
If you are interested in discussing this, or any of my projects with me, shoot me a message and I’ll get back to you in a timely manner.