Optimization

Viewing 10 posts - 1 through 10 (of 10 total)
  • #73045

    How would you optimize every variable including stop loss, trailing stop loss and or target – my computer would blow up doing this calculation. I assume you optimize indicators?

    #73060

    Either …

    Optimise  1 variables at a time
    OR
    Use very coarse / big steps in values and then narrow down successively
    OR
    Set opto running, go and dig the garden, swim in your pool (as Nicolas) or drink wine or beer (?) on your yacht (as Vonasi) 🙂 🙂 🙂

     

     

    #73063

    or drink wine or beer (?) on your yacht (as Vonasi)

    I have to drink pretty quickly as these days most of my optimizations are on daily, weekly or monthly charts without tick by tick and only one or two variables maximum.

    …..then again drinking quickly is my kind of challenge! 🙂

    1 user thanked author for this post.
    #73198

    Chris, im guessing u got the normal PRT account, so u can do maximum 2 variables at a time and very few combinations?

    if u get a premium account theres a limit on like 100K combinations which is… alot.. and u can enter all ur variables if u’d want to.

     

    if u only got the normal account id suggest optimizing “Filter variables” together, “Trigger variables” and “stop/loss”

    #73202

    This is how i work:

    https://www.prorealcode.com/topic/how-i-create-strategies-and-optimize/

    Look at how easy it is to curvefit, even tho it was profitable many years in a row, where i stopped optimizing (obvious what data i use to optimize lol, first 50% of the data) it turned out to be shit.

    #73245

    Yes I have a standard account is Australia – which is limited. Hopefully IG make me a premium client once they have made their $$$$ off me.

    I see up to 2009 you were killing it and then it turned on you.

    I have read your attached beginners guide and found it invaluable.

    #73275

    Yes I have a standard account is Australia – which is limited. Hopefully IG make me a premium client once they have made their $$$$ off me.

    I see up to 2009 you were killing it and then it turned on you.

    I have read your attached beginners guide and found it invaluable.

    Cool, glad u got something out of it. And ye the algo worked good until it turned on me completly. And that happens when u curvefit. If u keep it simple, dont have too many rules u should be fine even on new data. Another way to give ur algo green or red flags is to check it in similar markets like other indicies (dax, nasdaq, spain, italy, nikkei, hang seng, wall st, us 500 all look alike, even tho its a difference in the value of the index so u might wanna just remove/change ur stop loss and target, but i like to just run the exact same algo in all those markets just to see what would happen.) Spain is probably the hardest test of them all cus its seen some shit, and with premium on spain 1h+ u get data going back to 2001 i belive. so u 17 years of data is nice to have. Not saying ur algo should be perfect in every market, but if u make a strategy for wall st and it looks shit in ALL of the above ^ im gonna guess its curvefit and i would very vary of running it live.

     

    Btw about premium,. i started out with normal IG account, then i read on this forum that if u open an IG account thru PRT u get premium for free (or same cost as thru IG, amke 2 trades pr month and its free) so i opened an account in my SO’s name and got premium. Very well worth getting if ur serious about this. Its gonna be very very hard not to curvefit your models with only 2-4 years of data in my opinion. Much harder than with 10+ years.

    #73289

    My strategies are very simple indeed. I just use 2 or 3 emas depending on which works best. Maybe to simple?

    I have asked what other indicators i could consider to use but no replies as yet unfortunately.

     

    #73354

    I have asked what other indicators i could consider to use but no replies as yet unfortunately.

    In the Library and the ProOrder Support Forum there so many strategies using so many Indicators I can’t keep up! 🙂

    Also you may be like me and can’t understand lots of the  complex code?

    If I slow down and choose one System which at first glance I think … yeah I understand what is happening here … then try this and that and Graph a few variables, the understanding grows?

    Maybe try above and let us know?

    #73355

    My strategies are very simple indeed. I just use 2 or 3 emas depending on which works best. Maybe to simple?

    I have asked what other indicators i could consider to use but no replies as yet unfortunately.

    probably too simple yes. Normally what u want is a Filter and a Trigger. A filter is trying to prevent u from taking alot of false signals. lets say Close needs to be above 200MA, if its below, it wont take trades, because u see from the backtest/testing that the majority of losses in ur trades happens when close is under 200MA.

    A trigger is when to buy, when close is above 200MA. Lets say when MACD crosses over signal line + close above 200 MA. Then u got 1 filter and 1 trigger. Ofc then u need exit, stop loss etc.

    also about “what to use” the answer is “everything”! Just try everything that makes sense and then some things that dont make sense. Try reading up on what all the indicators actually do and how to interperet them. Its even a description in PRT when u click a indicator. Every indicator is made by someone who tried to interpret something within the market. Theres no indicator thats gonna give u perfect signals but they might help u determine where momentum is going or coming from. Or when a mean reversion might happen.

Viewing 10 posts - 1 through 10 (of 10 total)

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