morning Ftse mib 30minTF trading strategy

Forums › ProRealTime English forum › ProOrder support › morning Ftse mib 30minTF trading strategy

Viewing 15 posts - 1 through 15 (of 18 total)
1 2
  • 04/10/2017 at 9:44 AM #31578
    Francesco78

    Hello everyone, I have come up with a strategy on Ftse mib, based on the idea to take the opposite direction of the market after a day in which the market has moved at least 150 points, with a vol filter because I have the impression the strategy work if volatility is not too low. I also added and antimartingala, increasing the position step by step if the strategy remains positive. I saw that many people are posting backtesting on 30 minutes time frame with more than 100, 000 observation. I was wondering if you can help me in extending the backtesting range in this case as it looks really cod the last 2 years but I dont think is enough to be confident to use it on real. The code is attached. Many thanks.   Francesc

    04/10/2017 at 11:00 AM #31591
    Nicolas

    Hi Francesco, I did backtest with this version of your MIB strategy, but I don’t get the same results at all. Even with tick or not, with/without spread.. what if your time schedule for this index?

    04/10/2017 at 11:17 AM #31595
    Francesco78

    Hi Nicolas, thank you for your answer.

    That is strange, I am relatively new to this platform but so far all the codes that I have downloaded from the forum gave me exactly the same results as represented by the member, so I’m wondering what could have happened in this case.

    Im using a 30minutes time frame for this backtest. The strategy is to buy or sell at 9 in the morning and close the position at 9.30. Im using Italy 40 DFB. I also tried with shorter timeframes and the results are consistents, although the filter are not acting in the same way for the 2 different timeframes.

    what do you suggest to do to reconcile our tests? Im really wondering what could be wrong and Im a little bit concerned that data in this platform might not be so reliable if we get results so differents.

    Many thanks

    Francesco

    Many thanks.

    04/10/2017 at 12:12 PM #31599
    Nicolas

    We got the exact same data, no worry here. What is your GMT timezone for this particular instrument? Right click on your chart select “custom trading hours/schedules”.

    04/10/2017 at 12:13 PM #31600
    GraHal

    I got good results, almost exactly same as yours  Francesco78 and the Walk Forward results are super also!  I’m well impressed!

    I knocked 1 hour off the times to get back to UK Summer time, but I’m sure Nicolas thought of similar?

    GraHal

    04/10/2017 at 12:23 PM #31602
    Francesco78

    Ok, good news the date are consistent, I’m using Central European Summer Time.

    Thank you GraHal,

    Have you had a chance to do the backtest with 200,000 observations?

    Thank you.

    04/10/2017 at 12:29 PM #31605
    Nicolas

    Thanks grahal, I’m currently on leave and have my brain resting 🙂 I changed the GMT offset and get the same nice result. Doesn’t look good though before 2013 but it doesn’t mean the strategy is not good for the actual times.

    I’m wondering about the spread for the MIB ITA40?

    What about the result with no reinvesting?

    04/10/2017 at 12:30 PM #31608
    Francesco78

    One caveat for this strategy is the following.

    The bid offer spread for Ftse mib shrinks in a period between 09:00 and 09:00:30.

    If you want to run this algo in real pls be careful that you need to change the position opening time at 09:00:30

    I left 09:00 only to be able to backtest it for a longer period.

    Thank you

     

    04/10/2017 at 12:33 PM #31609
    Francesco78

    Thank you Nicolas,

    The spreads on the index is around 6 ticks after 09:00:30

     

    04/10/2017 at 1:17 PM #31617
    GraHal

    If timeenter is changed to 09:00:30 then the bar from 090000 to 093000 will be missed altogether and zero trades will be executed?

    Current spread on my Platform is 8, but then I am on UK Spreadbet.

    Like you Francesco78, I can’t do BT on 200,00 bars, sorry

    GraHal

    04/10/2017 at 1:49 PM #31619
    Francesco78

    yes zero trade on the 30minutes time frame.

    One option might be to run the code on a 1minute time frame and set the enter time at 090100.

    Also, the vol filter must be adapted to the new time frame.

    My idea is for this to multiply by 30 ( if we switch from 30min to 1 min) the observation of the Average true range indicator, that is to say from 50 to 1500, in order to make the 2 codes equivalent in the 2 different time frames.

    Would be great to hear if you agree on that.

    Many thanks.

    04/10/2017 at 2:06 PM #31620
    Eric

    I wonder if you can open a trade a 0900 when using a market order?

    When using market orders on a 30-min timeframe prt check if the conditions is true at the end of the bar and open the trade at next bar open

    run it on demo then you know

     

    04/10/2017 at 2:16 PM #31621
    Francesco78

    Thanks Eric, good idea, will try this approach.

    04/10/2017 at 2:22 PM #31622
    Nicolas

    You should do backtest with constant lot size and with the method of IS/OOS I described in this post: https://www.prorealcode.com/topic/how-to-use-prt-walk-foward-test/#post-31329

    You’ll surely get benefit to develop your ideas this way 🙂

    04/11/2017 at 12:11 PM #31709
    Francesco78

    Thank you Nicolas, yes I actually did ( or I think I did), the results seems good as far as I understand, Im only a bit puzzled by the fact that the variables optimizations suggest to use a “small b” where b is the width of the previous day bar. Optimization suggest a value between 50 and 75 while in my manual optimization I found a value between 125 and 150

  • Author
    Posts
Viewing 15 posts - 1 through 15 (of 18 total)
1 2

Find exclusive trading pro-tools on ProRealCode Marketplace

Similar topics:
combining manual trading and auto trading
conversion of a trend following strategy from the multicharts trading software
high frequency micro trading strategy – aapl
my most profitable trading strategy >10k per week
2 trading strategies in one system
adding input from custom indicator into strategy
andrew cardwell strategy on rsi
array/bubble sort in strategy doesn’t work, but works as an indicator
auto restart automated strategy
automatic trading is not authorized for this instrument

Create your free account now and post your request to benefit from the help of the community
Register or Login