Not logged in
Viewing 2 posts - 1 through 2 (of 2 total)

Forums>StrategyQuant>General Discussion>Different view before process WFM

  • #270784 |
    33 Posts

    I’m Just provide a different think with this article title.
    Do you think WFM result important? If yes, what’s the different between running simple optimization or not before WFM?
    Let me explain it, we know we build the original strategy then put it to robustness test.
    If passed, maybe we will running simple optimization (or not).
    If yes, you will get the optimization strategy of orignial.
    So, at this monment we get 2 strategies with the same logic but different parameter(or same).
    Now, let’s do the WFM.
    You can see we need to set the “Value distribution”, up, down and step.
    So, do you think about it before? The range and step will different between original strategy and optimization strategy. Because they have different parameters start point(or same).
    There are 2 situations with it (if they only use move average) :

    1.Parameters is close to the original after optimization.

    2.Parameters is far away to the original after optimization.

    If we set “the same” ranking option in two different condition above,

    At first condition, we will get the similar WFM robustness score both on original and optimization strategy.

    But, if we test WFM at second condition, it will get different WFM robustness score both on original and optimization strategy.

    I think you can understand what I mean, we don’t know our strategy parameters similar or not after we optimization.

    But, we will use that same ranking filters test our strategies with WFM.

    It’s doesn’t matter good or bad.
    It just a little thinking in WFM.
    If you want to share your thinking, please, and thank you.

    788 Posts

    i dont like WFM, because every with every run you can get different results – sure it depends on many factors – how do you have set the WFM, how many params your strategy has, how many params you are optimizing which how many steps and what is your param ranges, etc. etc.

    and in the final you will get some set of the param you will use in your real trading – but those params can make from your original strategy totally different one, and if you run the robustnest testing on this changed strat, it could fail

    for me its more comfortable to discard strats which have died on unseen data and generate new one continuosly and not to make periodical reoptimisations

    You want to be a profitable algotrader? We started using StrateQuant software in early 2016. For now we have a very big knowhow for building EAs for every possible types of markets. We share this knowhow, apps, tools and also all final strategies with real traders. If you want to join us, fill in the FORM. 1500+ final SQX strategies for members running on demo accounts to verify the edge DEMO ACCS. We provide also strategies for indices - DAX, DOW JONES, NASDAQ, SP, UK, because we have realtick data from our brokers.

Viewing 2 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic.