Developed a very good algo... Now what?

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

  1. wintergasp

    wintergasp

    It's more about looking professional than the reality of tickstory / FXCM.

    The only platform you may use to look professional is Deltix but that's 10k$ per month per computer.

    Nobody serious has FXCM-based track-record. In the futures industry, TradeStation is fairly good, we use it at work often to try quick ideas, yet if you walk around with a TradeStation track-record you won't be taken seriously.

    That's the only reason why we pay 16,000 £ per month to clear at JP Morgan instead of 0£ per month to clear at Interactive Brokers.
     
    #21     Mar 27, 2017
    aex likes this.
  2. algofy

    algofy

    BSD
     
    #22     Mar 27, 2017
  3. Couple years ago I met a programmer who was on a team that designed and build the platform Deltix,still keep in touch with him.
     
    #23     Mar 27, 2017
  4. wintergasp

    wintergasp

    It's extremely powerful but a bit pricey IMHO, they have lost quite a few clients since AlgoTrader and a few other providers started offering similar if not better software at lower cost AND offering it open source.

    At some point they had something like 120 employees for 30 clients or something
     
    #24     Mar 27, 2017
  5. I was a bit critical and this is always frustrating.I will write this as another guy sitting next to you and speaking his mind.To give you good analogy about robustness and all that.To me this manifested itself in team sports.

    My moniker is number 33 and when i was watching sports and also analyzing semi random sets of data,be it any stock,commodity prices i was intuitively searching for all around good solutions.And this is difficult,because when i made one part better,other part suffered in performance.Strive for limited weak points i understood better before i knew definition of robustness

    Scottie Pippen was always my hero,not Jordan.I always much more admire all around solid stats.In approach to things this concept fascinates me.

    5.jpg
     
    Last edited: Mar 27, 2017
    #25     Mar 27, 2017
  6. I was looking at the linked timeframe (live) which is quite short.

    PS your embedded pdf charts are a very low resolution and you can see squat if you zoom in
     
    #26     Mar 27, 2017
  7. Just curious. Are you implying that IB is accountability incompetent? Confused here.
     
    #27     Mar 27, 2017
  8. Adibi83

    Adibi83

    Ok, last thing though: I'm not reaching 8 units in 2 of my algos; only in one. In the other algo I can get to 4 units and that's all. As for the one algo that can go to 8 units - not only that it's rare, I also took into account a "black swan" scenario in which I lose several times in 8 units. That's why I use low leverage in the begining.
    I think my system is quite robust, as much as you can use this word in the forex market. Let's say that if one they I will be in an 8 unit position and suddenly a violent flash crach will occure to the other direction and my stop loss order will not be implemented, that could be ugly. But that kind of scenario have the potential to crash many algos and brokers, as you probably remember from the SNB case.

    Many thanks!
     
    #28     Mar 27, 2017
  9. Adibi83

    Adibi83

    What do you consider to be long enough? 3 years?
     
    #29     Mar 27, 2017
  10. It all depends on the market, # of trades, the method used, the risk, too many variables to answer. A pro will know something good when they see it. If you're not getting responses, there's something they don't like.
     
    #30     Mar 27, 2017