Not logged in
Viewing 12 posts - 1 through 12 (of 12 total)

Forums>StrategyQuant>Application Support>*BUG/PROBLEM*Strategies on based on RENKO gives very different results between SQ & MT4 backtester!! MARK PLEASE HELP!

  • #114286 |
    Karish
    Customer
    430 Posts

    Hello again,

    first of all i searched the forum and googled it all before i posted here again,

    i have read this: www.strategyquant.com/forum/topic/3697-which-renko-bartype-to-use/

    after having problems with my barType settings at the AZ-INVEST’s renko plugin,

    and exported the RENKO history by the guide in AZ-INVEST with “barType=1”.

    now when i develop strategies based on RENKO, i get good results in SQ, but when i retest any developed strategy in MT4 backtester it gives very different results O_O.

    yes i did add the code lines as AZ-INVEST stated to the strategy code,

    this are the 2 lines:

     

    AT THE TOP: “#include <CustomChartingBacktest.mqh>”

    UNDER START: “if(skipFirstTickOnBacktest()) return;”

     

    im building strategies with 2.5 PIP spread + 1 PIP slippage + 1 PIP distance from price.

    setted the parameters of the EA at the MT4 backtester as it was tested in SQ, made sure i have the same settings in SQ and MT4,

    my broker have: UNDER 2.5 PIP spread + UNDER 1 PIP slippage + 0 PIP distance from price.

     

    ***And BTW i found it very strange that SQ finds only LIMIT orders based strategies and not even a single strategy based on STOP or MARKET orders.***

     

    MARK OR ANY ONE FROM THE DEV TEAM, PLEASE HELP!, i had a thread about this thing also here: https://strategyquant.com/forum/topic/3722-problem-with-renko-from-az-invest-i-get-no-answer-from-them-2-weeks-please-help-me-solve-those-things/

     

     

     

    i just dont know what to do, i used the CSV2FXT script and converted the data with barType=1, selected 2.5 and made sure my backtester’s spread setted as 25 ticks (my broker is a 5Digits)

    then it gave me an FXT file and CSV file for SQ, i generate strategies on this in SQ, and then when i test the strategies that are developed on this data in MT4 after making sure that everything setted up just like in SQ, and im testing in MT4 with tickstory with that same FXT file that the CSV2FXT has created for me (the CSV file for SQ made of that FXT file), I GET VERY VERY VERY DIFFERENT RESULTS, NOT EVEN CLOSE TO THE SQ RESULTS, NOT EVEN CLOSE! O_o,

     

    please could somebody help me out with this?, maybe it is something wrong that i’m doing, or maybe its a bug or a problem with SQ?

     

    i dont know what even the problem..

     

    tried different strategies and different spread and parameters when i was backtesting in MT4, but still it dosn’t come even close to the SQ’s results………….

     

     

     

     

    PLEASE PLEASE HELP ASAP………!,

     

     

    I ATTACHED THE MT4 FILE & THE STR FILE**

    *This is just one of my strategies, they are all showing good results in SQ but very poor and different results in MT4*

     

    http://www.filedropper.com/290920152ststrategy

     

    http://www.filedropper.com/2ststrategy

    #133285
    Moderator
    1559 Posts

    Renko + Limit orders + MetaTrader .. I think this will probably never match. Fills would be very inaccurate compared to live trading. I would always use Market order on bar close with alternative timeframes because they move with volatility so you are able to get a good price even with Market order.

     

    I never had success with robots on range/renko chart. But I daytrade them discretionary. I can try to look at it but StrategyQuant can work with other and better platforms naturally supporting Renko or rangebars like Ninja or Tradestation

    #133300
    Karish
    Customer
    430 Posts

    Thanks for the replay Tomas,

    ok an alternative for MT4 always sounds good :), i may consider that.

     

    so if i generate strategies on renko/range bars should i change the Selected TimeFrame Only to Bar Open Only?, because i when i generate i generate on Selected TimeFrame Only mode…. is that wrong?

     

    when you retest your generated renko/range bar based strategies in MT4 (built based on Bar Open mode?), do you have any different results?, or everything stays the same?

     

    thank you.

    #133304
    Customer
    197 Posts

    And you can download and backtest with ninja for free.

    Sent from my HUAWEI MT7-TL10 using Tapatalk

    #133318
    Karish
    Customer
    430 Posts

    i just noticed there is a huge problem with strategies that generated for renko,

    the results in SQ showing you all the trades list, it shows your the SL and TP profit/loss trades

    but when you put this strategy into a demo account

    you start noticing some weird trades that are closing in small losses without any reason and probably not even strategy based orders,

     

    i mean the SL and TP trades are exactly the same as in the trades list that SQ shows,

    but in when im running that same strategy on demo there are some losing trades that haven’t mentioned at the trades list in SQ, check my example below

     

    renko bars is very buggy and un-finished project in SQ, i would really want it to be fixed……

     

     

     

    what i mean is in the trades list in SQ it shows like this for example:

    +10

    -12

    +10

    +10

    -12

    and nothing else, just the SL & TP like it should.

     

    but in demo trading sometimes there are some unexplained trades that are not based on the strategy O_O.. for example:

    +10

    -2

    -3

    +10

    -12

    -12

    -1

    -3

    +10

     

     

    you see all these small trades are not shown at the SQ trades list when you generate a strategy

    but when you put it to a forward demo test, you then will see some weird and unexplained trades that are closing in minus without any reason O_o…..!

     

     

     

     

    please fix it those renko problems, it seems that the renko side of things in SQ is not even 55% done…, very disappointed.. hope Marc or someone could tell me what is going on with this, im having only a hard time and problems with renko in SQ…….., and im a renko guy so thats my main thing..

     

    I wanted to add that SQ’s MQL4 code should give out a log that shows why the order was opened/closed with the reason also,

    example:

    Order # was opened: RSI 20 > RSI 50.

    /

    Order # was closed: SL.

    Order # was closed: TP.

    Order # was closed: BreakEven.

    Order # was closed: TrailingStop.

    Order # was closed: # Bars passed.

    Order # was closed: RSI 20 < RSI 50.

     

    something like this…, thank you for your time, i really hope it would be fixed soon….

    #133321
    Karish
    Customer
    430 Posts

    please also check my last posts here too about that problem:

     

    ¢ MT4’s backtest wont show the same results as SQ with limit orders based strategies,
    i have not checked it with market orders or stop orders because SQ cannot find any strategy for me that is based on market orders as-well as stop orders…, only limit orders based strategies…….

     

    http://azinvest.freshdesk.com/support/tickets/1249

     

     

    Over all i have only 2 problems:

    the one i mentioned right now above here, and the one i mentioned at this post:

     

    https://strategyquant.com/forum/topic/3786-bugproblemstrategies-on-based-on-renko-gives-very-different-results-between-sq-mt4-backtester-mark-please-help/#entry13318

     

     

    *Im getting a lot of great results with SQ with renko

    but the MT4 backtester and forward demo test shows different resutls,

    the MT4 backtester shows different resutls with limit orders.

    and the forward demo test gives some strange not strategy based small lossing orders, the strategy uses SL & TP to close the orders but there are some small orders that closing without SL or TP and without any comment or reason.

     

    thats all, please fix it asappppp……..

    #133333
    Karish
    Customer
    430 Posts

    *****UPDATE:

    Tested with Market Orders, all went fine everything is the same between MT4 & SQ works perfectly,

    but then why Limt orders based strategies wont give the same results in MT4 backtest?…, i think Stop orders will be with that problem also…

    the problem not only when backtesting Limit order based strategies on MT4 and it will show different results, but theres another problem that when you put the strategy into a demo account, the strategy makes some strange orders that are not mentioned at the SQ’s Trades List tab of that strategy and those orders are closing in a small loss most of the times and not closed by SL/TP like it should be closed by the rules, SQ does not show when the orders closed by time also.. 

     

    please fix this problem ASAP…… thank you.

    #133389
    Moderator
    1559 Posts

    Renko + Limit orders + MetaTrader .. I think this will probably never match.

     

    Can you post trade lists (SQ + MT) where we can see “strange” trades clearly and strategy generated by SQ (.str)?

    #133404
    Karish
    Customer
    430 Posts

    At my first post i posted the STR and MQL4 files

     

    i already moved to generate only market orders with Selected TimeFrame mode.

     

    i gave a perfect explanation in all of my posts here about this problem..

     

    a strategy generated with Stop/Limit orders works nothing like SQ shows…

     

    it does not match if you retest with MT4, and the trades does not match when you put it into a forward demo trading it will give false trades here is another example for it:

     

    lets say you generated a strategy based on stop or limit orders, your exit need to be only by SL or TP, your SL is 30pips and TP is 20pips,

    at the trades list in SQ you will see that everything right and all of the trades exit by SL or TP,

    but when you throw it on a demo you will see some strange trades like so:

     

    -30pips

    +20pips

    +20pips

    -5.3pips

    -2.7.pips

    +20pips

    -30pips

    -1.3pips

     

    so as you can see there are some false trades that havent mentioned in the SQ side of things,

    you can try the STR file i uploaded at the first post and convert it to MQ4 and put it on a demo and you will see..

     

    i moved to using only market orders with selected timeframe only because of this problem…., that really sucks because it close huge doors for me now..,, hope it could be fixed somehow.. it only got 2 problems false MT4 backtest results and false trades when trading it in demo/live..

    problem is only with stop/limit, market orders works perfect, i wish stop/limit will work like this also..

     

    thank you for your time and replay.

     

     

    BTW: here is another thread talking about this: https://strategyquant.com/forum/topic/3697-which-renko-bartype-to-use/

    #133440
    Moderator
    1559 Posts

    Can post your renko price data if possible with instrument settings you used? And what Renko barType you used with CSV2FXT plugin? 1 or 2? You need price data for candles with ‘wicks’ i.e. correct high/low

    I will do some tests with renko tomorrow with different order types

    #133453
    Karish
    Customer
    430 Posts

    tomas, i deleted all the strategies i had using Limit orders, the results are very misleading between SQ & MT4 unlike Market orders which are absolutely fine in that case,

    i used and still using only barType=1 (Renko with wicks), when i generate the CSV data,

    yeah give it a go, start by testing the Limit orders first, then move out to Stop Orders (which in my opinion will give the same misleading results that Limit orders would give), and then try Market orders,

    you would see that only Market orders will give the same results between SQ & MT4, which is very odd to me and it is such a waste that i cannot use renko bars with Stop/Limit orders…., it closes huge opportunities for me and in whole.

     

    if it would be fixed i couldn’t be happy and thankful enough,

    thank you for your time and help.

    #135672
    Customer
    430 Posts

    I am new on this forum and new user of strategyquant ( 30min ago).

     

    Renko bars can not be accuratly backtested. 

     

    The bars  do not accuratly display true trade data for Open, High,Low & Close of each bar. So when saving the renko data the data you get is false, As you might know it also paints, fills all gaps with generated data.There is some special types of renko bars were you can turn of the wave mode wile backtesting and the data will be accurate. I have tried that and bought that bar type but what happens then is that the trades position changes due to that the indicators now are calulated on the new data. Totall waste of money. So I would try and not generate strategies based on renko bars or any other fancy bars it will not be accurate. 

Viewing 12 posts - 1 through 12 (of 12 total)

You must be logged in to reply to this topic.