Forex Blog

First-hand Forex trading experience and information about foreign exchange market that will be useful to traders

Archives

RTATC2011 — Day 8 — Forward-Testing and Final Back-Test

September 4, 2011 by Andriy Moraru

Almost one month has passed since I last posted about my Road to ATC 2011. Some important changes were applied to my MT5 expert advisor that I plan using during the contest. After a period of extensive back-testing the EUR/GBP pair was substituted with EUR/USD and Z-Score optimization was turned off on both pairs. Additionally, fixed fractional position sizing (in which position size growth with profit and declines with losses) was substituted with the flat but aggressive “3 lots per position” model.

The latest (and I hope, the final) version of my ATC 2011 EA can be downloaded for free:

The forward-testing was rather unsuccessful for my EA. Not that the results were bad — I’ve noticed that the forward-testing isn’t any different from back-testing for this expert advisor in MetaTrader 5 — but I failed to test it uninterruptedly for more than several days in a row. First, I had serious problems with the internet connection for days, then there were real-life issues that prevented me from keeping the EA running and, finally, there was a 3-day downtime period on MQL demo servers. But the main result of this handicapped forward-testing is the knowledge that the EA runs and trades smoothly. Additionally, a complete similarity between the back-testing and forward-testing results makes me believe that back-testing can be trusted this time.

Back-testing on the period between 2011-01-01 and 2011-09-01 — the period, which is used by the ATC 2011 automated EA testing system — ended up with $82,878.58 profit, which is quite modest compared with the top results of the previous Automated Trading Championships. Nevertheless, I am still satisfied with it. You can download the back-test report:

The main part of the development process actually ends here. But I’ll keep you informed of any dramatic changes in my EA before the championship trading starts.

If you have no idea what RTATC2011 is, I recommend reading this post.

P.S.: 18 days left for registration in ATC 2011. 1,831 participants registered so far. I bet the number will surpass 2,000 soon.

If you have any questions or comments regarding the last tests of my ATC 2011 expert advisor, please feel free to reply using the form below.

10 Responses to “RTATC2011 — Day 8 — Forward-Testing and Final Back-Test”

  1. FX Lover

    Hi

    I am wondering if there are blog ranking site for Forex in USA.
    Do you know blog ranking sites for Forex in USA?

    I appreciate it if you let me know it.

    Thank you for your support.

    Reply

    admin Reply:

    I’m not sure but maybe this is what you are looking for:
    http://www.blogcatalog.com/category/business/investing/futures/

    My site also has a rating of the Forex blogs (but not only US):
    http://www.freeforexsites.com/category/blogs/

    Reply

  2. martino

    go go go! i will follow your atc 2011 challenge and i hope you will win the contest!! and i know many others followers of your blog are thriving. by the way i appreciate you decided to turn off z-score, it seems to me some kind of over-optimization. ciao. martino

    Reply

    admin Reply:

    Thank you! Z-Score can be a nice thing if done properly. Unfortunately, it should be recalculated constantly within the EA itself and this is quite a difficult thing to work smoothly during the ATC.

    Reply

  3. martino (it's me again...)

    hi. i’m very curious to follow your challenge. just a question: why did you decided to abandone the fixed fractional position sizing in favour of the flat 3 lots model? in my opinion the fixed fractional scheme prevents from risk of ruin and allows the exponential growth of the equity. in fact the backtest with the fixed fractional approach is largely better than the fixed lots one. is there something i don’t understand? i hope you’ll be the winner of 2011 atc!
    martino

    Reply

    admin Reply:

    I’ve tried backtesting with different position sizing approaches and flat model gave the best results. I know its disadvantages, but fixed fractional showed much worse results. That’s probably because the system relies on rare larger wins, usually followed by losses; so the fractional sizing significantly increased losses.

    Reply

    martino (it's me again...) Reply:

    thank you for the fast answer. i assume that is not correct to compare the backtest you posted in day 6 with the one you posted in day 8. thank you again, i will follow the competition…
    martino

    Reply

    admin Reply:

    It’s not correct because the latter is two months longer.

    Reply

  4. martino (it's me again...)

    ok, thank you for the explanation. let’s go and beat all the others!

    Reply

    admin Reply:

    Thanks! I doubt that I will beat all the others, I just hope to end up with profit.

    Reply

Leave a Reply

required
required (will not be published)
optional