Automated Trading Developer

Discussion in 'Automated Trading' started by WS_MJH, Apr 10, 2012.

  1. WS_MJH


    Wondering if anyone has a good recommendation for an automated trading developer or development firm. I do not code myself, and I'd be looking for a firm to help with the research and development of the system. There are lots of hucksters of there, so if any of you guys have any tips for success, I'd appreciate that as well.

  2. Are you expecting to pay the developer?

    Or would you be expecting the developer to accept a share of any upside?

    What platform do you want code developed for?

    Or are you agnostic about platform at this stage? In which case, which broker are you using (which will determine which platfroms might be used most easily)?

    Or are you also agnostic about the broker?

    Where are you located, and will the location of the developer be important to you?
  3. WS_MJH


    Re: payment. Of course I'll be paying. As for upside, that would be more of a partnership. Thus, the characteristics of the relationship would be much different.

    Re: platform, broker. I currently use thinkorswim. Since tos doesn't have auto trading, I have to go to another platform. I know most of them and know most brokers, so I'm agnostic on that front. Location does not matter; I'm not really in a trading, tech metro area.
  4. Mr_You


    Unless you find a partner, and a partnership doesn't look very attractive IMO, I doubt you'll find ANYONE who is willing to help you with R&D.

    You're better off downloading NinjaTrader or using TradeStration/MultiCharts and finding a programmer to code the strategies and indicators per your request. But keep in mind this can be frustrating for both the trader and the programmer: vague or impossible requests, bugs, unrealistic expectations, etc.
  5. I have to be honest with you here. Any coders worth their salt with the experience required to build a robust algo trading platform aren't going to be cheap. Here in London you'd struggle to get a good one for less than £500 per day. A great one will cost you much, much more.

    You'd be able to pick up a decent coder without algo trading experience for a lot less but then they'd be learning the skills on your clock and coin.

    How much you got to spend?
  6. WS_MJH


    This is one area where you don't get cheap.
  7. Div_Arb


    TOS offers automated trading via their Prodigio platform. If you use indicators for your strategy, it's pretty good. If you use price action, it's not so good.
  8. MattZ

    MattZ Sponsor

    I have a few questions:
    How many conditions you estimate are in your logic?

    When you say development, are you looking for a semi automated, fully automated, and do you also require help with the risk management components.

    We have worked before with a number of good developers, but I noticed that less complex methods are better on EasyLanguage while more complex ones are better on C#. I might PM you once I have more details for someone who I consider professional.
  9. WS_MJH


    Div_Arb -- re: prodigio. As you may know, tos is getting rid of prodigio. I think it's totally gone in a month or so. I tried using it. When I installed it, I got nonestop bugs and memory problems. After some time with technical support, I got rid of it and reinstalled tos. prodigio is a good concept; auto trading for people who don't code. But the whole system is sort of clunky. I do think tos might in the future do an auto trading system with thinkscript. I have no idea if and when that'll happen.

    This was a great comment. It was sort of what I thought. I read somewhere easy/powerlanguage can produce a concept with less codework, but it just didn't seem as powerful as c#. Also, you'd be stuck with multicharts/tradestation, while with c# you could move onto more professional platforms.
  10. Not sure if your concept is tested or not, but remember, it is easier to prototype/back-test with a platform like ninjatrader/multicharts (EasyLanguage) than in C/C++.

    If you method is already working, go for C/C++. If not, test it in an already existing platform first. It will save you time (and money).
    #10     Apr 10, 2012