Beginner seeking guidance

Discussion in 'Automated Trading' started by GlobalMacro, Aug 3, 2014.

  1. Hi All,

    I have been reading these forums for quite some time, but I decided today was the day I would start posting and contributing to the EliteTrader community.

    Currently, I work at a small global macro fund which uses a discretionary strategy. I love the discretionary strategy very much and has done wonders for our group, but in this central bank controlled environment I want to begin something different and start a more systematic approach in investing.

    I have zero programming experience and want to build pricing models and automated strategies but do not know where to begin. I was wondering if you guys could give me some guidance and steps to begin this journey.

    First of all, how can I begin to learning programming, and if so which languages are the best in this field? Which "IDE" should I also use? I honestly do not know what's next after learning a language.

    Appreciate your help guys as always and hope to here back from you soon.

    Thanks,

    GM
     
  2. xandman

    xandman

    Welcome.

    Perhaps, you should choose a language that is compatible with your firm's systems and methods.

    You mentioned Global Macro. A lot of econometrics models? R and Matlab is Excel on steroids.

    Since your already in the discretionary trading camp, no need to put in 4,000 hrs to become a great programmer. Choose a fast prototyping language such as Python and have a pro code it into other languages when needed.

    Seems the likely pairing for interoperability of language( a higher level goal) is R+ Python or Java/C + Matlab..
     
    Vinayak Maheswaran likes this.
  3. Couldn't agree more with Xandman. That is all the programming you really need. Try to network with programmers who can help you with what you really need to know as opposed to having to sort through a whole lot of unrelated material.
     
  4. I do not agree with the suggestions. The OP said he knows no programming. I am surprised that anyone would suggest Python to a non-programmer. His way to go is with something like easylanguage, powerlanguage or Amibroker AFL. Since when is prototyping with Python fast? Let us provide an example, a system that goes long at next open when today's close is higher than the high of 5 days ago and reverses to short when the close is less than the low of 5 days ago. This is the basic EL code:

    If c[0] > h[5] then buy Next Bar at open;
    If c[0] < l[5] then Sell Short Next Bar at open;

    Let us help the OP. I would like to invite xandman to write the same code in Python since he suggested it.
     
    Last edited: Oct 8, 2014
  5. xandman

    xandman

    I'll try to stay on topic without swinging my d*ck around. The guy works in a professional trading environment. If he wants to stay relevant or progress, he has to go with what the industry demands.

    Check out the job listings for python in Chicago or go thru the hiring page of several Prop firms.

    http://www.simplyhired.com/k-python-developer-trading-l-chicago-il-jobs.html

    For 99% of traders, you will make more money in an institutional setting. I am not knocking TS or Multicharts, mind you. I use Ninjatrader myself.
     
    Last edited: Oct 8, 2014
  6. Mr Glo mac;
    Be sure to paper trade your ''automated ''for about 10 years;
    that way your discretion maybe will be improved-NOT a prediction. Simply a system comment.Wisdom is profitable to direct.In other word like Rich Dennis said,'' its the job of the market to wreck computer systems''
     
  7. Nobody cares if you do have a d*ck and what you do with it. You may go to a porn forum to talk about that. But some care if you have actually used what you suggest. You can only prove that if you provide the Python code for the two lines of EL code I posted.

    There are no industry standards in trading. I worked for two large hedge funds in the past that used EL exclusively. If you cannot make money with EL or AFL, you will not make money with any other language.
     
  8. I agree 1000% with you. I'm a former chief architect of a fortune 50 company with huge experience under my belt in object languages, although I have never programmed python, and understand it is the preferred language by many trader developers for its rapid development features, plus there are libraries that help you shorten development. There are also similar capabilities with Ruby, in fact Tradier API is written entirely in Ruby. Its much more approachable than python, I'm surprised more people haven't latched on it.

    But I wouldn't for a second consider diving into, and trying to create my own trading platform when you have a mature platform like Tradestation. Easylanguage is much more approachable, and quite powerful. I know many professionals that use it, including myself, and have run a hedge fund strategy with it. The institutional support at Tradestation is excellent.

    Don't let the name EasyLanguage fool you into thinking it is underpowered, it is not. You can also develop libraries and extend the language if you are that ambitious...but someone starting with no programming experience will have a much easier time with EasyLanguage...pun intended.
     
  9. One problem is that the world is full of programmers that seek to transform everything into a programming adventure. A way to create a market. The simplest systems are the best and EL or AFL are more than enough to do the job. Someone I know still has systems in Metastock formula language that work well. I understand the need of using such languages to develop complicated algos for other things but for trading of the type done by 99% of the people in this forum Python is an overkill and a waste of time.