Stock-Prediction-Models transaction costs and slippage

Hi @huseinzol05 Is it possible to add transaction costs and slippage into the back tests? Many thanks, Best, Andrew

Aug 31 '19 10:08 andrewczgithub

I will do it for only realtime agent. Comparison among 23 agents just to study capability to trade naively

Aug 31 '19 10:08 huseinzol05

Hi @huseinzol05 Ah amazing. Sounds awesome! How are you going to model the fees? Many thanks, Best, Andrew

Aug 31 '19 12:08 andrewczgithub

@andrewczgithub I am still struggling connecting with websocket as the websockets fetches data in real time and in json. I have converted the json strings into array and try to feed them into neural net one by one . But failed. Another approach we have tried to saved it onto DB but failed . Can you guide us what's the right method and format of data to put into neural net from websocket. we have make sure our data is in ['close', 'volume'] dictionary. Have you been suceded in connecting to websockets aor any guidance to do it

Sep 01 '19 01:09 augmen

I will just do simulation from 3 years historical data CSV. Not connecting any api / websocket.

Sep 01 '19 06:09 huseinzol05

That would be awesome! I was actually just looking at adding in a commission to the live agent somehow last night but didn’t want to break it haha. Some kind of editable transaction fee, and an explanation on how to add more features than just Close and Volume would be perfect! I’ve had lots of fun and see a lot of potential in your work for real world application 👍 keep up the great work. We all appreciate it!

Sep 06 '19 19:09 windowshopr