Quant analytics: Artificial Intelligence meets Artificial Stupidity

(Last Updated On: October 6, 2011)

Quant analytics: Artificial Intelligence meets Artificial Stupidity

Rather than clog up other people’s discussions with this little debate, lets see if we can keep our voices down and have an intelligent discussion.

Can neural networks make money in the markets? How about “biological” artificial intelligence? What if you add a random number generator into the mix? What if you remove every known indicator too


I would be satisfied if mr.Writer will publish his screen shot to simplify the contest


What are all those screenshots supposed to prove? We’re in Automated Trading Strategies remember. Screenshots aren’t much use to automated trading software.

I wasn’t “in” today. That’s because only a part of my TA trade log for yesterday looks like this –http://screencast.com/t/dzi7UQUUz


These are the screenshot of the systems that are trade automatically…

Apropos similar conversations elsewhere, to me they seem more appropriate for the “Promotions” section.

Is there any chance you could take a look at Holger’s links and let me know if your methods might possibly be squeezed into that sort of framework? As you must have realised by now, I am rather keen on “backtests”.

I’m rather keen on “forward tests” too. I like to load the software under test onto a VPS, set it running on a demo account, then take another look in 3 months time to see how much pretend money it has made (or not as the case may be!). At the moment I can’t do that with your software either. My screenshot reveals one of the numerous reasons why that is the case. Is there any chance you can fix all the “bugs”?


You were in great help!
I’ll let you know.


Pardon my shortness here. I am establishing a connection to this thread. Thank you.

Welcome to a novel variation on the stale old “backtesting is BS – Oh no it isn’t” thread. As you can see, things have been a trifle quiet in here recently, unlike some other places one might mention!

To restate the question posed elsewhere, can you see any way of shoehorning your methods into the sort of framework outlined?

You are referring to the now apparently “missing attachments”? They were there 7 months ago!

I wonder where they’ve disappeared to? I’ll see if I can track them down. In the meantime hopefully the general gist of the thread is apparent. Are your methods amenable to “objective” comparison with a variety of potential “black boxes”, based around use of a common, agreed dataset?


In my opinion, I think any challenge such as this needs to be broken into 3 portions.

1) Feature selection Challenge
2) Preprocessing Challenge
3) ML design Challenge

Participants should submit 3 distinct API’s which will be compatible with the winner of the previous stage.


The short answer to your question is yes. I have been a practitioner in the computational intelligence field for a decade now.


Do you guys by any chance fancy participating in the design of an “Automated Trading Challenge” of some sort. Maybe even entering the competition itself?




NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!
Don't miss out!

You will received instantly the download links.

Invalid email address
Give it a try. You can unsubscribe at any time.


Check NEW site on stock forex and ETF analysis and automation

Scroll to Top