Not logged in
Viewing 11 posts - 1 through 11 (of 11 total)

Forums>StrategyQuant>General Discussion>Optimal settings for strategy generation exist. Its not personal preference.

  • #113588 |
    Customer
    484 Posts

    It is a quantification. It doesn’t matter what indicators or settings you “like”.

    I’m starting a thread to compare settings and to search for the best. I’ll update my settings tomorrow and logic. I highly doubt they are the best but why do I think there is a “best”?
    1)Filter out BAD indicators (yes, indicators can be “bad” and are also not a “preference” when it comes to quantification). We want robust indicators that most often produce robust strategies that last the longest and don’t blow up accounts 5 months in.
    2)Understand that SQ has s settings and indicators that are not beneficial to optimal robust EA production means you can randomly generate more strategies faster with higher likelihood of meeting your filtered requirements.

    With a thread sharing reason’s WHY we use certain indicators, (or all?), certain filters, certain settings, we can begin to find the most logical choices. When we filter out has many settings as possible, we will generate more strategies that meet our criteria.

    This is especially important for anybody with slow PCs!! (not me)
    Even with a fast PC I still don’t want to waste time generating settings that are not beneficial hence why this thread is being started (probably the most important thread to date, and its a mystery why there isn’t one with consistent postings).

    We are quants. Remove the unnecessary and increase odds of finding optimal strategies.

    I’ll update some of my filters and settings and indicators I remove for generation tomorrow.

    Edit: I made it in video format.
    In brief:
    https://youtu.be/ji4HArEPEjg

    Comment here. Share your ways too. This is just mine. Part of doing this is to start a discussion and have other SQ users input their logic of doing things.

     

     

    #129803
    Customer
    3 Posts

    As a newbie to SQ I have spent the last few days looking up the definition of all the indicators MT4 and SQ, and the ability their ability to work in conjunction with each other, and the combinations and possibilities are enormous…

    If you hit the holy grail, I am not asking for your find.  However, listing which combination of indicators you used that had poor results are appreciated to eliminate the unnecessary.

    #129845
    Customer
    484 Posts

    I relpied to this a couple days ago but it kept saying “saved for later”.
    Any way to retrieve to post?

    #129846
    Customer
    484 Posts

    updating this weekend

    #129910
    Customer
    484 Posts

    Will be recording video of my settings and putting on youtube tonight. I’ll link it and make brief explanation of some of them in text here.

    #129914
    Customer
    484 Posts

    In brief:
    https://youtu.be/ji4HArEPEjg

    I’ll add some very brief written reasons for my choices in this thread in near future, basically what is said in the video.

    Comment here. Share your ways too. This is just mine. Part of doing this is to start a discussion and have other SQ users input their logic of doing things.

    #131545
    Customer
    19 Posts

    Thershold,

     

    I am surprised that no other users replied to this post except very few and kept the momentum going… thanks a lot for sharing your video.. I am new to this and been using SQ for past 2 weeks and still struggling to find good set of indicators/filter criteria. Debating between random vs. genetic.. but i am determine to find best settings by investing my time and doing trial and error.

     

    i will post my settings the moment i like the equity curve of my results… but i will share it for sure…

     

    For now i am only focusing on Crude oil and US indexes.. (5 min, 10 min , 1hr chart ).

     

    I wish other experienced user should share their knowledge about criteria( how to avoid bad ones, what are the best once they like etc…) within building blocks and ranking options… It will significantly improve other users turn around time…But again it is a personal choice of individuals and i will completely respect that…

     

    i believe personally in knowledge sharing so at least from my parts, i will continue to ask questions and share my good findings.

    #131551
    Customer
    484 Posts

    I may remake this video in the future and will certainly add more eventually.
    Gentmat has recently shared more information on what causes the most time consumption in the past couple days than I’ve ever seen anyone else share. I learned quite a bit from a few of his recent posts.

    Thanks for the reply. Glad it was helpful.

    #131566
    Customer
    19 Posts

    Yes future video will help for sure. I went through Gentmat post about Zulu and memory settings.. 

     

    i just applied those settings and now running it with new settings. However i have low configuration in my PC so setting suggested by Genmat will surely speed up the process

    #132019
    Customer
    26 Posts

    Yes, this thread is very helpful.  Thank you.  So far I failed to create a SINGLE strategy in SQ on a set of data that can be easily modeled in TradeStation ( though the model is admittedly poor, but it is matter of a little work). I deliberately picked a set that has well formed trends and sideways periods  Your video about settings helped me to pick up some information that is either absent in manual or I just missed it.  

    #132020
    Customer
    557 Posts

    Hi Threshold,

     

    I just watched the first video and the amount of similarity to your approach to mine is quite reassuring. Many thanks for posting.  I know you posted it a while ago, but I don’t have much spare time these days lol

     

    I will comment on a few things that might be useful to the discussion.

     

    • Candlsticks – agree, only make sense on weekly and daily timeframes if anything.  Below daily I agree, worse than useless.
    • Heiken Ashi – agree, never seen these produce anything and one strategy that did was completely different in MT4, so maybe bug in SQ or MT4….
    • General Indicators – almost the same as you, quite weird really. I like to select the simplest things, like moving averages and oscillators, and avoid everything else.
    • ATR for SL – I find the ATR very laggy, which means it chooses small stop loses when volatility is increasing and hangs onto big stop losses when volatility has paused, so tend to select fixed pips.  I think something is missing in SQ, which is SL based on other indicators such as fractals or recent support and resistance levels.  This is where a large fixed stop loss followed by trailing stop in SQ work well at the moment.
    • Test precision – I tend to run in tick simulation for anything below M15, just because a lot can happen in one bar regarding hitting stop losses.
    • Selection Criteria – very similar to me.  I give stability  5, use net profit and profit factor, ret/dd, minimise stagnation.
    • I used fixed lots, because it’s easier to see what is linear and stable and compare strategies. 

     

    I also use random for long periods.  Genetic produces “samey” results I don’t think it adds anything.

     

    Thanks for posting.

     

    Mike

Viewing 11 posts - 1 through 11 (of 11 total)

You must be logged in to reply to this topic.