This algo is just me trying to learn how the IDE works and learn what I can and can't do. The algo takes 20 stocks ranked by the highest 30-day momentum and runs the stocks through an ARIMA model and computes a forecasted value. If the forecasted value is greater than the most recent close, long the stock, and vice versa. Putting this up to see if anyone else can add something better to it or tweak it so I can learn.
As it stands now, the model is extremely computationally expensive, and the backtest runs very slow. The way I currently have it set up, it rebalances and runs the ARIMA model every single day, which is why it is so slow, and why I limited the pipeline output to 20 stocks. Please give me any suggestions or ideas you guys have so I can see how to make it into something useful.
