V12 – Odd or Annoying Happenings!?

Forums ProRealTime English forum ProRealTime platform support V12 – Odd or Annoying Happenings!?

  • This topic has 113 replies, 11 voices, and was last updated 1 month ago by avatardruby.
Viewing 15 posts - 91 through 105 (of 114 total)
  • #238015

    Roberto do you not get problems with the optimiser re-optimising variable values 1, 2 and 3 where what you want to optimise is variable values 4, 5 and 6.

    I’ve used variables values 1 to 6 to steer clear of describing variable values as sets, as that is what PRT call them in the setting that you say (above) you do not use.

    Hope that makes sense.

    #238022

    Yes, I occasionally get weird results, as well. It’s difficult to tell where the problem can be, since it seems the source is not always the same.

     

    #238024

    If we use the ‘SETS’ then the weirdness (optimiser re-optimising variable values just set as fixed) ceases / is no more.

    But the thing is, using the SETS stops weirdness that was never there in v10 (v11 also?) and so I cannot see why PRT introduced the SETS??

    And very quickly I end up with 30 to 40 SETS (on one System) which I never again use and which – to delete – requires me to delete each SET individually.   100’s of mouse clicks on < 10 Systems … to achieve what??

    Trouble is I am getting other weirdness, almost like a System is picking up the wrong SET of variable values … hence why I am considering deleting the 100’s of redundant SETS … but I could do without the Faff / admin task and I might even get RSI from all the mouse clicking in such a short time! 🙁

    #238025

    Had a look around, couldn’t get the error message you showed.

    Only things I came across was, If you choose the top,(default) option in the ‘save result’ window, it creates a new trading system which appears in the ‘Indicator & Trading System’ windows list. This will be different to the one(s) you ran.

    The other option creates a new ‘set’, with, displayed settings, which appears in all the set lists.

    However, doesn’t appear in  ‘Definition of variables’ window set list, if its open. Appears when re-opened.

    Regarding the “easy way to delete” all the sets…

    If you choose the top option in the ‘save Results’ and create a new system file, then there won’t be any optimised variables and only default set.

    In the new file, all the original variables appear at top of code file with a preceding ‘ONCE’ keyword which can be easy cut from code.

    If next question is, that’s good but I want to keep the optimised variables’ , I only can think of save a file version before creating all the sets.

    I suppose if you have a version with just optimisation variables and default set, then revert to that when you want to dump all the sets.

    I know I’m simplifying  it a lot, and more likely the process is a lot more messy and doesn’t lend itself well to the simplification.

    But,  as my boss says, ‘Adapt, Modify and Succeed’, the other thing he say’s is ‘Yes’ a lot,  and when you ask how!, he reply’s Somehow!

    regards

    1 user thanked author for this post.
    #238027

    Thank you for above druby , I’ll try your suggestion. I do save all my .itf files on my PC so I can always revert.

    What is annoying though is that PRT have come out with this SETS function which nobody yet has stated how it can be used usefully / help us with optimisation etc.

    We are having to spend time discussing how to get around / nullify what (presumably?) PRT thought was a good idea??   But PRT forgot to tell us how / why it is a good idea?? 🙁

    If we don’t use the SET’s then we are beleaguered by the optimiser re-optimising values that we have just entered as fixed in the Definition of Variables Table.

    #238028

    I think ‘Set’ is kind of miss leading even though it fits. (a set of Settings)

    All the variables placed in the list, appear in all the sets, and its only the ‘Value setting’ that can be set. (fixed or optimised)

    So in each set you can only alter those options and the values.

    The info syntax is    var[fixed]   for fixed and    var[min,max][step]    for the optimised.

    If more than one set, options appears in right pane, unchecked defaults to default set, checked gives set option and info on active set.

    The set chosen in the right pane overrides the one in left

    For Automatic TS, the edit boxes display the vars fixed value(s) of the active set. Default is ‘1’ unless they have been changed.

    Or can be set  to other values manually.

    These end up at the top of  ATS file as   ONCE VAR = FIXED / MANUAL from my observations.

    From a work flow point of view I can only think of this off top of my head…

    1. Add all variables to list as fixed with user fixed settings
    2. Choose a var to optimise and set it optimise min/max/step
    3. when optimised report comes up, save as set with relatable name using best result.
    4. change to new set and edit, update and fix first vars new value, and optimise var2 with min/max/step
    5. keep repeating till all var fixed with optimised values.
    6. Save as new trading system

     

    #238029

    @PeteSt, I’m sure you love them and can enlighten us all how we could put the ‘variable sets’ to good use?

    Hahaha ! Sure !!
    Probably …

    I think you refer to what we talked about the other day (3 weeks or so back). But do you ?
    For now I don’t understand the max of 5 vs the 31 you show. Is that the same subject ? “Hundreds of times” what ??
    Maybe you found an Easter Egg ?

    You must be talking about what I show below – is that right ? Well, if so, the meaning of your text in English is .. eh .. wrong ? and the meaning of what I show is understandable. Mine is a noun (the parameters to optimise over) and yours is about a verb of some sort (optimise). And your list of 31 (34 ?) … I don’t recognise that.
    Didn’t someone promise release notes for this century (hopefully not from the past century) ?

    I can lay out some good use of those sets later, but I first like to understand what you are looking at, and how ever you can come up with 31 (34) of them. Something seems to be not right. And a max of 5 ? I think I use 7 (but not sure). This is in V11.

    #238048

    SS1

    Screenshot1 (SS1) shows the Warning about 5 Sets of variables which I have now exceeded (I can’t find where they are to to delete them!?).

    We can save a Set of variables in a ‘new’ System by clicking on the icon 2 columns to the left of the warning message in SS1.

    SS2

    SS2 shows 34 Sets of variables which I have optimised for one System.

    Each time we change the layout / values of variables, we have to add a new number or description to make a new Set of variable values.

    If we do not do above, the optimiser re-optimises the last values optimised, the result is we feel we are losing it, cannot believe what we are seeing, feel demoralised and overall very frustrated with PRT Platform v12.

    I have about 15 Systems now each with 30 to 40 Sets (none of which I need / want!)  … hence about 500 clicks needed to delete about 500 Sets!   In a few weeks, I could easily need around 1,000 clicks … you get the picture.

    I’m going to start deleting the previous Set each time (so then the numbers of Sets will not build up) but it all feels like faffing for no good reason as v10 was not set up lik this. I need to check v11 as memory dims?

    #238062

    Screenshot1 (SS1) shows the Warning about 5 Sets of variables which I have now exceeded (I can’t find where they are to to delete them!?).

    It seems that you are saying that it is only a warning ?  I mean, how else can you end up with 34 of them ?

    Otherwise, deleting is done here (see below).  But highly annoying is the fact that the first set never can be deleted. I if I have created a nice (thus additional) set which is the only one I want to retain, then this is not possible and I always will have two at least, of which the first one I will never use. And this is highly dangerous when copying the while Strategy because the selected one is not pertained in the new copy (both/all are still there, but the preferred selection is not).

    #238064

    how else can you end up with 34 of them ?

    The 34 are Sets of variable values for optimsing.

    The Warning that I have  exceeded 5 Sets is > 5 Systems created / saved using the icon (2 columns left of the warning message in SS1). 

    So – despite, confusingly using the term SET – the saved System Sets > 5 is not to do with the 34 Sets of variable values in the Definition of Variables (DoV) window.

    It’s best, for now anyway, that we limit discussion to the 34 Sets in the DoV window and how you use them usefully?

    #238065

    After trying different things and combinations, I eventually got the error message. image1

    Could not isolate down why, or re-create it again, yet!.

    The error message came up when I pressed OK to create a new ‘set’ using the ‘save result’ option in the ‘Optimize Report’.

    At this point I only had the default set and 1 variable, but I had been creating and deleting  sets, variables, trading systems all from the one original backtest file.

    I did find other  error message if you click on a result line ‘To backtest with tick by tick data’ as commented at the bottom of ‘Optimized Report’, and you have deleted those variable and/or set, out of the ‘Definition of variables’ list. image2

    1 user thanked author for this post.
    #238068

    Use case of the Sets (my idea of that) :
    (not really how you use it 😉 )

    Along the way I have normalised all program code plus their Parameters, so that one code can serve one physical market. Example : All what is Financial Future. So that Nasdaq, Russel, (Eurex) Dax, Stoxx and a few more, are served by one code. This is physical too because it is one Strategy (one itf when saved).
    The Strategies are made unique by means of their parameter settings. Thus, Trailing could work differently for S&P vs Nasdaq vs (for sure) DAX. This is thus all differentiated in the Parameters and the code obviously is consistent with all the parameters.
    A new Future (e.g. DJI) should work with the same code but with new Parameter values. Maintenance is now suddenly doable.

    Precious metals require a different base code, but Gold vs Silver expresses by Parameter set.
    Precious Metals Futures again require different code (but mostly the same – hey, where are these due Functions hanging out ?!?).
    Forex Spot will again bear very different code, but all Forex pairs should work with the same code.
    Currency Futures will again require new code.
    Etc.

    What does NOT work in this (ahem) concept is that production versions require the proper Set selected. This NEEDS to be saved in a separate .itf, or else mistakes will be luring. Once saved in an .itf, the selected Set will be saved with it (thus, which one was the active one, is saved with it). Thus, restarting the Strategy will not require the reselection of the Set (but I always check it).
    What runs where and when (or was stopped) is maintained in a redundant list (per account) of which you see a snippet below. Without that list, it is impossible to deal with it all. But on-topic : besides the name you also see the Set. The name (of which I promised it was general like “Financial Futures”) is made dedicated to the saved production version because else I won’t know what runs where. Remember, the code of all of them is the same. The set is indicated in the list too, like Std (for Standard first Set) or QO (for in my case “Quick Out”).

    Another dimension lies in the fact that the code may change in areas where the Parameters do not. Typical example would be the new covering for a (newly found) broker anomaly, such as not trading at Broker reset times (like 06:30 Amsterdam for IBKR). This will create a new code + name, but is unrelated to the Parameters hence how the particular instrument behaves for Trailing in relation to the Timeframe used.


    So as you have seen I do not use the Sets for backtested versions. Maybe some times I do that, but it is way too dangerous because it is too easy to select the wrong one and before you know it you work with the wrong set and head a wrong direction without you realising.
    This is personal of course.

    All summarised : I use the Sets only to have one code for the same “physical” market (e.g. Financial Futures). This is really worth while for maintenance.

    1 user thanked author for this post.
    #238070

    The error message came up when I pressed OK to create a new ‘set’ using the ‘save result’ option in the ‘Optimize Report’.

    Ahaaaaa … So that is what GraHal is doing too !
    Right ?

    I have never used that. And I deliberately did not. I would be detached from all reality. Read : I need to backtest my “set” version and I do that with the one Strategy (code) and the one (or two) Sets in there.
    Again, this is personal.

    #238071

    So that is what GraHal is doing too !

    Yes, but I only did it as a trial, thinking, mmm could this be useful?  Wish I hadn’t bothered!!

    The saved System then appears in the Trading Systems and Indicators (TS&I) window.

    I have deleted the saved System(s) referred to above and do not have any Systems showing in the TS&I window, but I still get the warning message telling me I need to ‘delete some existing sets of variables’.

    #238077

    the saved System Sets > 5 is not to do with the 34 Sets of variable values in the Definition of Variables (DoV) window.

    Wrong again GraHal! 😀
    I duplicated the System and then deleted 34 Sets of variables (so back to default Set only) and the Save function (via icon in the optimise results table) was restored for this  particular System.

    One mystery solved!

    1 user thanked author for this post.
Viewing 15 posts - 91 through 105 (of 114 total)

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