My first post of hopefully more to come!
Please note, I have no formal TA training and the lines are just a rough guide to get my main SMA crossover point across. Please also forgive me for my vague phrasing.
A friend and I have been researching and testing a few different neural network models, LSTM and reinforcement. We are now getting deeper into the LSTM model to predict near term price changes, later incorporating market sentiment (positivity) that we will hopefully use to construct a portfolio of non correlated instruments that will be automatically readjusted frequently to push fund allocation towards positive predictions and away from negative ones.
Now to this chart, while we were looking at different models and backtesting strategies we came across a simple program where you can search for the optimal (most successful) lead/lag SMA periods for crossover trading. While this doesn't mean that the crossover periods will continue to perform as well or at all compared to the backtest, it has worked the best so far.
I'm going to be making a small entry if the lead crosses above the lag by 2.5%, with selling to be done when it falls below it by 2.5%
Let's see what happens!
Please note, I have no formal TA training and the lines are just a rough guide to get my main SMA crossover point across. Please also forgive me for my vague phrasing.
A friend and I have been researching and testing a few different neural network models, LSTM and reinforcement. We are now getting deeper into the LSTM model to predict near term price changes, later incorporating market sentiment (positivity) that we will hopefully use to construct a portfolio of non correlated instruments that will be automatically readjusted frequently to push fund allocation towards positive predictions and away from negative ones.
Now to this chart, while we were looking at different models and backtesting strategies we came across a simple program where you can search for the optimal (most successful) lead/lag SMA periods for crossover trading. While this doesn't mean that the crossover periods will continue to perform as well or at all compared to the backtest, it has worked the best so far.
I'm going to be making a small entry if the lead crosses above the lag by 2.5%, with selling to be done when it falls below it by 2.5%
Let's see what happens!