New Trade Strategy

Discussion in 'Automated Trading' started by Mins, Jun 27, 2007.

  1. Mins

    Mins

    I have recently nearly finished my trading strategy, it is a strategy for trading YM futures.

    The equity curve is attached and some breakdown on returns is also given.

    I am trying to tweak it to smooth the equity curve. Is this a decent equity curve with decent percentage profitability.

    I am new at trying to automate strategies and was hoping to send the orders via my esignal to IB TWS but keep having problems with the API.

    What is the best automating trading software out there??

    Any advice and feedback would be welcome.
     
  2. lrm

    lrm

    What did you use for your back testing?

    Thanks,

    Louis
     
  3. Is the scale correct in your posted equity curve (it starts at $0)?
     
  4. Mins

    Mins

    Here is some trade analysis:
     
  5. Mins

    Mins

    I used esignal for backtesting and it produces all the details for me.

    I then went through a few trades to make sure it was not bugged but it seems legit.

    I dunno though because i am quite new to this.
     
  6. Your curve and the time span of the back-testing looks curve-fitted. What happens if you test over a longer period or forward test in the future?
     
  7. Mins

    Mins

    My settings for initial capital was $2500
     
  8. Mins

    Mins

    I cant get esignal to give me any more back testing at the moment.

    It has been tested on various timeframes but the 7min fits the best.

    It averages +60% winning trades on nearly all time frames i have tested.

    Like i said i am new at this so i may have made a mistake somewhere.


    Any help on best automating software, best broker automate through???
     
  9. That's truly phenomenal Mins!

    OVER $225,000 in 1.4 weeks. It looks like your (backtested) starting capital was quite small, too.

    Is this possible?

    Ga
     
  10. lrm

    lrm

    Well, from what I can tell this equity curve says that you turned $2,500 into more than $200,000 in 7 days.

    I don't want to sound like a naysayer, but something smells fishy here. One of the most common problems in back testing are a class of errors known as postdictive errors. These are subtle logic errors in which you use a future event to determine what to do in the present.

    For example:

    if price(bar+1) > price(bar)
    buy at market

    That is an obvious postdictive error. It checks if the next bar is higher than the current bar, and if so it buys at the current market, i.e. the current bar's price.

    Good luck!

    Louis
     
    #10     Jun 27, 2007