Developed a very good algo... Now what?

Discussion in 'Automated Trading' started by Adibi83, Mar 26, 2017.

  1. algofy

    algofy

    Possibly look into a dynamic margin levels instead of static. I add and subtract size/margin relative to current account balance and recent history. Just a thought if you don't get an investor.
     
    #11     Mar 27, 2017
    Adibi83 likes this.
  2. Keep running it. Don't increase the risk. Don't meddle or change it.

    Finally you're unlikely to get interest from hedge funds based on a myfxbook account.

    GAT
     
    #12     Mar 27, 2017
  3. Adibi83

    Adibi83

    So how can I get their interest? The myfxbook account is synced with my FXCM account... it's real.
     
    #13     Mar 27, 2017
  4. Adibi83

    Adibi83

    I thought of that but after several attempts I realized it's just does not fit any of my trading strategies
     
    #14     Mar 27, 2017
    algofy likes this.
  5. These days money goes rather in characters than in systems.Do you have charisma,can you juggle or dance tango(whatever comes to mind...)? Good luck!
     
    #15     Mar 27, 2017
  6. wintergasp

    wintergasp

    1) FXCM is sub-retail. I'm afraid you've lost your time so far. Your best bet is to open an account with IB as it's probably the only discount broker that is considered serious by smaller institutions. Same goes for myfxbook as @Adibi83 mentioned

    2) You'll need a better presentation than this, I see 5-10 hedge fund presentations a week and they all look the same, you need to look the same if you want to be treated seriously.

    3) Nobody likes FX when it comes to serious investors, your best bet would be to try to adapt it to the futures market or cash equities. The mindset behind this is that we don't see any big FX fund that actually work but see many equities and futures fund that work well, so nobody will risk their clients money on a FX manager (remember nobody invests their money, you always talk to someone who manages the money for someone).

    4) Don't ever say "reliable back-testing" with MT4 screenshots. Try to build your own back-testing engine in C#, that is really the base of quant trading and will allow you to improve your approach much faster
     
    #16     Mar 27, 2017
    aex and Xela like this.
  7. I am sorry to say,but your method is not robust at all.
    You Martingale adverse signals to the point that you think is safe and then there is cut off point.
    This tells me you set fixed these parameters based on median point,yes backtest.
    Remember you biggest gain is also your potential biggest loss so going from 1 unit to 8 is a big NO NO.And this is a problem,because you think you have a solution to this issue by running 5 strategies just to smooth the equity curve resulting from equity drawdown when you double your bets.Running remaining 4 strategies cover for that some of the time.
    Where you place your stop loss on 8 units is not indication of good risk control.The whole idea to get to this point is flowed from robustness standpoint.
    I would not invest with you.
     
    #17     Mar 27, 2017
  8. Adibi83

    Adibi83

    Thanks for the reply. However:
    1. "You Martingale adverse signals to the point that you think is safe and then there is cut off point" - if that was true you would have seen much more risky trades of 8 unit leverage, and not only 2 risky trades since 2004. The fact is that risky trades are rare in my system.
    2. "Remember you biggest gain is also your potential biggest loss" - you're absolutely right, and I took it into account when examine the biggest potential loss that can occure. That is why I always start at zero to low leverage.
    3. "going from 1 unit to 8 is a big NO NO" - again, you are right. but I never go from 1 to 8. Let's say that a lot needs to happen until I get to 8 units. If you want I can send you the full results of my back test so that you would be able to see the trades. These HTML files are heavy, and I can send them to you via wetransfer.
    4. "you think you have a solution to this issue by running 5 strategies just to smooth the equity curve resulting from equity drawdown when you double your bets" - I don't double my bets all the time. I have 5 algos and 3 of them won't double at all. never.
    5. "remaining 4 strategies cover for that some of the time." - and what's wrong with that? isn't that what they call hedging? When you go long on a stock and buy some PUT options for defence - is that wrong too?

    I honestly appriciate your opinion and advice. Thanks again
     
    #18     Mar 27, 2017
  9. Adibi83

    Adibi83

    Thanks. My comments:
    1. Yes, IB are better than FXCM, but the fact that my system is profitable under FXCM as my broker (with 2.5 pips per trade, whereas IB gives 1.5 and lower) says a lot doesn't it?
    2. Could you please send me one to take a look? I will aprriciate it alot (adibi83@gmail.com)
    3. Futures. Noted.
    4. Tickstory isn't reliable? And again - my system is not day trading or something like that. Even if tickstory's data if offset in 5, 10 or even 20 pips - it would not make a differnce.
     
    #19     Mar 27, 2017
  10. If it helps you get to another "level" it's all good,by this i mean going further.

    My view all along is you only double up adverse signals in one of the strategies.
    This is why i wrote "running remaining 4 strategies cover for that some of the time".
    To me this was obvious,sorry i did not put this in wording in first paragraph.

    I know you will not double up in all of them,because i done tests,manually to that so there is huge amounts of work done.

    Now you tell us you double up in two of them,at least cut this down to one if you have to.That would be risk control.But the system is not robust and nothing will cheat this important factor.I understand are going back to initial amount and hedging,nothing wrong with running 4 strategies that are not correlated to the one that exposes you to most risk.The problem is this rare situations when you put yourself to the wall and bet 8 units.Your weak point,because you don't know and nobody knows for sure the distribution of this occurrence.So far you run into two of them.If you run into few of them in short period of time you will not recover from this drawdown.This is what i am trying to say.

    You will go to another level if you are not stuck in a rut,and another level and another
    To find robust strategy is a large number of trials and errors.
    Good Luck !
     
    Last edited: Mar 27, 2017
    #20     Mar 27, 2017