I’d love to bounce ideas around, but I have to say that building a daytrade model is next to impossible, because of all the unpredictable factors that change from day to day.
A good daytrader enters his/her day already knowing what the news of what trade war is going on, what country is shooting missiles at what other country, is the Fed Chair giving a speech that day or not…..etc. Can’t create a model that predicts random occurrences that can’t be known. If Wal Mart announces blowout results on the day Trump announces new Chinese tariffs…….WalMart stock is heading down.
Now, if you increase the time horizon to, say, a month…..that mitigates what cannot be known, and allows you to model longer term trends like unemployment, consumer confidence, etc., and how those might affect trade over that month.
Sure you don’t want to model climate change? It’s probably simpler. :-)