Hello all. I don't have training in data science, and I want to know if I'm on the right path with my thinking for the algo I'm developing.
I wrote a strategy and tested it in TradingView. For many stocks I get a percent profitability score between 65% and 88%, but for some stocks it's even lower and the strategy loses money.
Next, I identified two variables I can use to improve the performance, and this can turn most losing stocks into winners. However, I don't yet know what factors determine the input for any given stock. I'm simply testing all logical inputs until I get the highest percentage profitability.
So I'm thinking I have two options.
- Either I determine the inputs on a per stock basis by running an automated backtest which finds the optimal value through iteration, or,
- I do some kind of factor analysis which will hopefully show me a relationship between the variable inputs and observed variables like market cap, sector strength, market bullish/bearish, float, etc.
What is the best path forward from here?
Submitted October 18, 2020 at 06:53PM by flakyfacefool