Is the SWAT worth being sold on prorealcode Market place ?

Forums › ProRealTime English forum › ProOrder support › Is the SWAT worth being sold on prorealcode Market place ?

Viewing 15 posts - 31 through 45 (of 50 total)
1 2 3 4
  • 01/30/2023 at 9:21 AM #208723
    Nicolas

    Best to do is to GRAPH your returned variables from the CALL on the price chart and compare with the indicator values applied to it.

    It is indeed possible that your indicator start to compute before the backtest so reults may differ depending of the way the calculation are made. Try to play with PRELOADBARS.

    01/30/2023 at 12:01 PM #208744
    LucasBest
    Nicolas wrote:

    Best to do is to GRAPH your returned variables from the CALL on the price chart and compare with the indicator values applied to it.

    Bonjour Nicolas,

    I did, and i can say from that, that there is a bug while using call and MTF. You can see the proof here below…

    Maybe it is due to the fact my indicator returns 15 variables ? Maybe it is too much ?

    First picture is the indicator alone on 1 hour time frame showing that it returns a signal the 15th of july at 5 am.

    Second picture shows that the bot don’t see any signal when running in 5 minutes timeframe. The indicator is called in 1h timeframe with updateonclose (but i tried also with default) and it seems that my indicator don’t send any signal… So the bot don’t do anything in this period.

    Third picture show that the bot works perfectly with 1 hour timeframe (for the code of my indicator) and 5 minutes timeframe (for the managment of position code), when i put the code of my indicator directly inside the proorder code (without using the instruction Call)

    The fourth picture shows that if i switch my graph from 5 minutes timeframe to 1 hour timeframe (which not allow me to keep the instruction Timeframe (5 minutes, default)) in the code of proorder) and change the timeframe insruction to Timeframe (60 minutes, default) then the proorder code with Call of my indicator works fine but i loose the benefit of MTF and i loose the ability to exit my positions more accuratly…
    Note that if i change the code Timeframe (5 minutes, default)) to Timeframe (60 minutes, default)) but i keep the default time frame of the graph at 5 minutes, it does not work either… So The Call instruction of my indicator need the main time frame of the graph to be set to 1 hour if i wanted to behave normally as it works normally in 1 hour time frame.

    So the conclusion is that Call and MTF have a big bug here, smetimes behaving normal and sometimes not : sometimes the signal does not appear, and sometimes new signal (that does not exist in 1 hour time frame) appear while usinf call preceeded by Timeframe (1 hour , updateonclose) but in a different main time frame on the graphic

    01/30/2023 at 12:07 PM #208750
    LucasBest
    Nicolas wrote:

    It is indeed possible that your indicator start to compute before the backtest so reults may differ depending of the way the calculation are made. Try to play with PRELOADBARS.

    I tried also to add preloadbars and If with enough barindex to execute, but i prevent only some false signal but did not correct all.

    The signal of the 15 of july at 5 am that i talked in my previous message does not appear either… The pictures i posted was done after this add in my code.

    01/30/2023 at 12:10 PM #208752
    LucasBest

    Also, i launched yesterday my bot in my real account with all the code in proorder :
    – 3000€ start capital
    – 12% risk at each wolfe waves as i am very confident on the bot behaviour
    – MTF with 2 minutes execution

    01/30/2023 at 1:00 PM #208758
    PeterSt

    Lose remark : I would never do something with BarIndex unless I very thoroughly had checked how it calculates under a TF setting.
    Or maybe the question :

    What do you expect from BarIndex under the 1 hour setting vs the 5 minutes TF setting ?

    <hr />

    Otherwise it still would be so that Indicators can’t deal with MTF. Or … prove me wrong ?

    All right. I just tried something myself and “something” unexpectedly works. Now :
    I did not hear you telling about the TF in your Indicator(s). But did you try to solve your problem with *also* setting TimeFrame commands in there ?
    Thus, in the Indicators you can work with TF commands just the same. I just did something which “relates” to the external TF in order (the one of the chart) and next mimicked what my calling code does for other TFs. This seems to “match”.

    I know, for many this won’t be an eye opener, but at least for me it is new that “this” could be done. Maybe what(ever it is) I do is new because it still is so that in V12 something has been created newly regarding TF and Indicators.

    Lucas, you should try too focus on :

    1. Your Return command in the Indicator is at the lower TF (5 minutes or 2 minutes);
    2. The math you do in the Indicator is – where required – at the higher TF (1 hour).

    Notice that it is allowed to have several A, B, A, B, C, A, B, A same TF’s switched in the code. This may be needed.
    Always explicitly start with a TF command of the TF of the chart (no matter it is automatically there).

    01/31/2023 at 9:25 AM #208804
    LucasBest

    Hi Peter,

    when i try to put the MTF in the indicator, it does not work either… The indicator close and it write “calculation error”.

    Strange that it works while the code of the indicator and the MTF are both in proorder code…

    This is weird!

    01/31/2023 at 9:28 AM #208805
    PeterSt

    Hi Lucas – Can I somehow mimic your error without unveiling your code ? I feel this is a matter of understanding how things are “calculated” and people (e.g. me) might be able to recover from that (or avoid that issue).
    Anyway the calculation error might be indicative for the issue in the base. At least I hope it is …

     

    02/05/2023 at 5:28 PM #209149
    LucasBest
    GraHal wrote:
    GraHal wrote:

    Show the equity curve with positions under the equity curve and price curve under the positions.

    I might have missed it, but what position size are you using?

    Why don’t you show positions under the equity curve and also price curve under positions?

    I finally understood what you were exactly expecting… At least i hope 🙂

    This is with 3.33% maximum risk for each wolfe wave that is traded : from 10.000 to more than 2.500.000 after 10 years.

    I realised also that my position size was calculated only with the maximum risk for each wolfe wave, but without checking if available equity would be enough to buy the amount of shares calculated with maximum risk per trade. So i added with lines of code to allow the bot to buy only the amount of shares it can buy with margin requirement… This made the results drop from previous position sizing, but i improved the bot and how it manage each wolfe wave and now it reach even better results even if the position sizing is correct now.

    I don’t know if you can see something from the equity curve with positions under the equity curve and price curve under the positions, because the equity curve is literally exponential!

    From IG Website :

    Stock index margins

    Margin requirements for CFD positions with non-guaranteed stops are capped at the amount of margin for no stop (ie if the stop is wide then the calculations used may give a higher margin requirement than the calculation for no stop. If this happens then we limit the margin to the amount required for the same position with no stop).

    The larger figure of the two calculations below:

    Number of contracts x contract size x stop distance + guaranteed stop premium
    Number of contracts x contract size x price x margin percentage

    02/05/2023 at 6:53 PM #209154
    GraHal
    LucasBest wrote:

    exactly expecting… At least i hope

    Yes, thank you.

    I can now see a position size of 2,500 contracts on US Tech 100 giving the big spike of over 2.5M early in 2023.  Clever that the strategy exits at the top so quickly … will allow you to sleep at night! 😉

    Cheers

    02/05/2023 at 7:20 PM #209156
    LucasBest

    The position size increase as the equity increase until the max of tier 3 which is 3000. After the maximum of tier 3, the margin percentage rate increase to 15%. But Tier 4 is not in the code yet.

    I still want to see how the bot behave in real life. But it is coded in a way it can’t loose more than the maximum risk you choose per wolfe wave, even with 2500 shares, unless there is a gap down (i imagine).

    Anyway, if this bot can do as good in real life as it does in all the backtests, i would be glad if it is able one day to buy 2,500 contracts on US Tech 100… Really!… 🙂

    For the last Wolfe wave it does not really exit at the top, as US Tech 100 is still going higher now. I am trying to improve the bot while target 8 is reached and when price go higher… Maybe a kind of trailing stop when last target is reached or something similar.

    02/05/2023 at 8:18 PM #209160
    GraHal

    We are mostly ‘little fish‘ on here 🙂 so if you are considering selling the strategy on Market Place then it be more realistic to start at position size = 1, up to max 10, maybe 20?

    I could not sleep with position size = 2,500 – even if based off Gains – what if there is a massive spike down and IG is NOT accepting Sell orders!? 🙁 … it has happened more than once! ).

    02/05/2023 at 9:05 PM #209161
    LucasBest

    Well, we will see this later, as i can’t sell it in the Market Place for now anyway… as i have not found any solution to the problem/bug i have while trying to hide the code by using a hidden indicator.

    So, i will just test my bot like this, because i don’t care if i loose 3000€ of starting capital if there is a chance that it can turn more than 1 millions 10 years after.
    For those who want only small gain, it is possible to set maxpositionsize to 1 or 2 or 10… But for me it is just wasting all the quintessence of this bot, which have twice bigger average gain than the biggest lost, and can make exponential growth of the start equity : because maximum lost per trade is contained while in certain circunstance the bot can double the capital in a single wolfe wave. Risk reward is simply amazing

    02/05/2023 at 10:28 PM #209165
    GraHal

    Please run your System on Demo-Live data before running on Real-Live.

    Do you have a Demo Account?

     

    02/05/2023 at 11:14 PM #209168
    LucasBest

    No, i don’t have. What is it ?
    And why should i run it in demo live instead of real life ?
    What can happen ? Can i loose more than the start capital ?

    I’am new in coding and using trading bots, even if i do technical analyses for more than 13 years

    02/06/2023 at 9:23 AM #209198
    GraHal
    LucasBest wrote:

    What can happen ? Can i loose more than the start capital ?

    If you ever get to pos size = 2,500 on US Tech 100, you would need 2,500 x 625 = 1,562,500 as Margin in order to even place the trade.  This is assuming you are a Retail Trader at 5% Margin.

    Surely you not want to lose  1,562,500 … even if you already have it??

    (Hope my coffee has worked through and my maths is correct?) 🙂

    Such things, you will not find out by backtesting. You need you a Demo Account in order to demonstrate how it all works.

    You need to enable your Demo Account, it is a setting in on the IG Platform if I recall correctly.

  • Author
    Posts
Viewing 15 posts - 31 through 45 (of 50 total)
1 2 3 4

Find exclusive trading pro-tools on ProRealCode Marketplace

Similar topics:
buy on certin dates, not on technical singal
exchange time on market
is it true that you can’t invest in eu stock market shares with autotradin
is there a easier way to hide the code on algos?
questions on market data
trailing stop on ib – stop repositioning on dax

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