This is a guide of the trading strategy. For a guide on how to read the graphs on the analysis page, click here.
(This strategy is still under experiment. It will be available once it is ready.)
Prophet is a state-of-the-art machine learning prediction model published by Facebook. It is used by Facebook to predict many aspects of it’s own internal systems, including traffic on it’s websites. You can find out more here.
In this strategy, we assume a ranging market. First, we fit the Prophet model to our historic data. Then, we use Prophet to predict today’s price based on the model and compare that to today’s actual price.
A BUY signal is generated when today’s predicted price is greater than the actual price. Exit when a SELL signal is generated.
A SELL signal is generated when today’s predicted price is lower than the actual price. Exit when a BUY signal is generated.
Prediction models like Prophet are very powerful because it is forward looking. However, these models requires fitting and the fitting process can be unstable. As a result, the prediction can fluctuate significantly. Hence, strategies based on these predictive models should be used with caution.