Not logged in
Viewing 4 posts - 1 through 4 (of 4 total)
  • #113083|
    Customer
    473 Posts

    Hi all,

     

    I’d be grateful if you could try to answer the following questions:

     

     

    Q1:   Is it possible to calculate custom indicators using range bars in SQ?

     

     

    Q2: is it possible to perform a WFT of a portfolio of instruments as opposed to a single instrument? 

     

     

    Q3: if I perform a WFT using RET/DD as the fitness criteria, for each WFT run does SQ simply select the optimized instance of the original strategy that has the highest RET/DD?

     

     

    Q4: Regarding robustness testing, when would it be appropriate to use randomize trade order with replacement?

     

    Thank you 

    #128247
    Administrator
    1177 Posts

    Q1: you compute custom indicators in MT4, so if you are able to run EA on range data then you can also compute custom indicator values

     

    Q2: not right now, we’d like to add it to the future version

     

    Q3: yes, it choses the parameters that have the best performance criteria (in this example Return /DD) in the previous optimization period

     

    Q4: what do you mean exactly by using randomize trade order with replacement?

    Mark
    StrategyQuant architect

    #128257
    Customer
    473 Posts

    Thank you.

     

    Regarding Q4: I referred to the monte carlo analysis (robustness testing).  One of the methods is to randomize trade order using either the exact or resampling (replacement) method so I wondered what are the advantages of exact vs resampling?   When is it appropriate to use the resampling method.

     

    Thanks again.

    #128298
    Administrator
    1177 Posts

    it is hard to say when is is appropriate to use which method. Each gives you little bit different image.

     

    Resampling is more “tough” on data, its results possibly have little bit more value.

    Mark
    StrategyQuant architect

Viewing 4 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic.