Strategy Eur/Usd mini, how to reduce the maximum % drawdown?

Forums ProRealTime English forum ProOrder support Strategy Eur/Usd mini, how to reduce the maximum % drawdown?

Viewing 15 posts - 16 through 30 (of 55 total)
  • #87409

    I made an MTF version out of latest Vonasi’s version, by:

    • replacing empty line 5 with timeframe(Daily,UpdateOnClose)
    • replacing empty line 48 with timeframe(default)
    • appending Nicolas’ Trailing Stoip code
    • launching it form a 2-minute TF

    as from attached screenshot.

    2 users thanked author for this post.
    #87415

    **** Google translator*******
    Thank you all for your contribution.
    In theory, the strategy should appeal to all markets with appropriate optimization, in fact I would like to create a diversified portfolio on this strategy.
    Roberto hrazie for your MTF code but are looking for a strategy that can betray in real and I think the MTF is not yet supported in real mode, or something has changed? Currently I use PRT of IG which has some limitation.

    #87417

    You are correct that MTF is still only in beta testing but we should all be using this opportunity to live forward test MTF strategies so that we are ready and confident in them when it finally gets put live. Also it helps PRT/IG find any bugs which can only benefit us too.

    Regarding your topic title and trying to reduce draw down. I personally think the draw down is already pretty low on your strategy and it will be difficult to improve on it without adding new conditions which will then just lead to less trades and less profit and increased chance of curve fitting. I usually have in my mind that every time that I add a condition that has a variable then I am multiplying the curve fitting probability by two. My theory is that if we keep it simple and it works then it is more likely to work going forward than a complicated strategy is.

    1 user thanked author for this post.
    #87418

    Thanks Vonasi
    in fact, my goal is to find a simple strategy that is stable.

    #87424

    Drawdown can be further reduced by applying the same strategy above to Heikin-Ashi candlesticks (still Nicolas’ trailing stop code uses regular japanese candlesticks).

    I tested it on regular EurUsd, rather than the mini one because I found out it is not just a simple math issue (it’s, or it should be, just a division by ten), because on the mini contract I’ve been reported more than 300 trades, almost all losing money. While there’s no difference when applied to DAX €25 and DAX €5, it’s the same behaviour, only multiplied/divided by 5!

     

    1 user thanked author for this post.
    #87465

    Ok grazie Roberto, lo proverò ma ho notato dal tuo screenshots che per alcuni mesi non scambia.

     

    #87473

    coscar – English only please in the English speaking forum – even if you are both Italian!

    #87476

    Sorry Vonasi, it’s the fault of the automatic translator.
    “Ok, thanks Roberto, I’ll try it but I noticed from your screenshots that for a few months it does not exchange.”

    #87479

    Hello Vonasi and Roberto I still ask you a kindness, you could test the attached code since 1995. Thank you

     

    #87483

    It is not possible to test it from 1995 as you need tick by tick data which is only available from mid 2010.

    #87485

    OK thanks, you could check the robustness of this code, I would like to start it in real time. Thanks for your availability

    #87487

    I have not had a chance to test your code yet but the only way to truly test robustness is to live forward test a strategy in demo. Personally I would not put any real money on any strategy until I have seen months of forward testing in a variety of market conditions and preferably lots and lots of trades. It is boring and you need a lot of patience but losing money is easy enough so why be in such a hurry to do it?

    I see that you have added some form of martingale and money management. Don’t do this until you have forward tested the strategy with level stakes – it just makes you imagine all the money you could be making when in reality you don’t even know if level stakes would make you any. Martingale is very dangerous unless you have a very high win rate strategy or one where the wins are much bigger than the losses and even then it can all suddenly get very expensive in just a few losing trades. If you have Warren Buffets bank balance behind you then use martingale – if not then don’t.

    All just my humble opinion but then people have all sorts of opinions on how to best lose money at this game!

    #102093

    I did a multitimeframe version of this fine system that seems to perform quite well in backtest, I haven´t tried it yet in real or demo

     

     

    #102096

    Next autumn v11 is likely to be made available to IG customers, thus making 1M bars available for backtesting.

    #102107

    Here is the code from the version posted by TempusFugit. My main concern if I had coded it myself would be the exit conditions based on days of week and time and bars since trades opened which seem to smell horribly of data mining.

    The true test is to remove most of them and see how it fairs. With just the exit if it is 2100 hours and the rest of them are removed then the equity curve is a little less inspiring. Image attached.

    If we let them back in one by one then the equity curve starts to look nice again which is really telling us that they are just data mining conditions. IMHO.

     

Viewing 15 posts - 16 through 30 (of 55 total)

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