Quantitative Analysis of Economic Calendar - New Free Site

[andrewstevens]

Hello I'm an active forex trader that heavily uses quantitative analysis and machine leaning models to make trading decisions from economic news data. I spend a lot of time researching new quantitative techniques some of which I have decided to share on a new economic news website that offers predictive analytics. Can I please get some constructive feedback on the site in terms of understandability and how it can be improved. Note this is a free site I don't want anyone's money.

Please let me know what you think of it

http://www.newscaptial.com/

 

[FXX]

Hello I'm an active forex trader that heavily uses quantitative analysis and machine leaning models to make trading decisions from economic news data. I spend a lot of time researching new quantitative techniques some of which I have decided to share on a new economic news website that offers predictive analytics. Can I please get some constructive feedback on the site in terms of understandability and how it can be improved. Note this is a free site I don't want anyone's money.

Please let me know what you think of it

http://www.newscaptial.com/

Looks interesting

What's your success with it?

 

[andrewstevens]

I've been generating positive months - that's with my additional experience as a trader. The automated backtest showed very good performance, but I find using the tool with my experience I get best results live. A model AUC of 0.6 is reasonably good, I will weight trades with a high AUC say 0.7 as the signal is more predictive. You can try it on a demo account and see how it helps you.

 

[FXX]

I've been generating positive months - that's with my additional experience as a trader. The automated backtest showed very good performance, but I find using the tool with my experience I get best results live. A model AUC of 0.6 is reasonably good, I will weight trades with a high AUC say 0.7 as the signal is more predictive. You can try it on a demo account and see how it helps you.

I have moved away from AI for my stuff as it isn't consistent and has margin of error. Its also somewhat annoying to have to build training sets for a mediocre result. I have therefore moved to my own algo which has thus far proved to be far superior than AI and comes with far greater control and flexibility. would be good to hookup as this area interests me a lot.