Anyone program their own systems/program?

Discussion in 'Strategy Building' started by swoop[TR], Nov 4, 2002.

  1. CalTrader

    CalTrader Guest

    Looked at your site markethackers.com : Doesn't inspire my confidence. Its not even finished. If your site isn't complete then why would I assume your software systems are ? Distributed systems are indeed not slapped together. Automated decision support / pattern matching / AI-like systems are very difficult to get right.

    Do you have some hard statistics that validate the performance of your systems ?

    Audited results would insprire more confidence. There are a number of big players in this space: a couple of groups around los alamos that I know. Please post your URL showing third party audited results and then I would be more interested .....
     
    #11     Nov 5, 2002
  2. You're right, the site isn't finished, but the system is. The data (candidates/open/closed positions) is constantly being updated in real-time throughout the day. I only posted the url to drum up interest and receive some feedback on our day to day results.

    You bring up some valid points; I will immediately look into providing third party validation of the system. Please continue to visit www.markethackers.com; we have backtested our trading system extensively and think we have a lot of potential.

    Thanks for your comments CalTrader.
     
    #12     Nov 5, 2002
  3. Can anyone here suggest a way to get study results i.e. stoch,MA,etc onto an Excel sheet so I can base my entry/exit off of those values. I currently have an automated order entry/exit system but I still have to plug in the buy/sell candidates. Before I had this system, I had a filter in Investor RT which spits out a list of buys/sells, I then key in order manually. Now it is the other way around, I can auto the orders, but would like to load the list onto Excel where my auto orders emanate. any insights? Thanks
     
    #13     Nov 5, 2002
  4. jaan

    jaan

    yes.

    yes (been live for 2 years).

    we did experiment with nets and GA, but eventually found them to have serious curve-fitting problems. at least in our particular case.

    - jaan
     
    #14     Nov 5, 2002