Trading System Building Learning Curve?

Discussion in 'Automated Trading' started by thefuturestrader, Apr 28, 2014.

  1. I am looking to start learning how to build trading systems. I think I want to start with basic trend following systems, like a bollinger band or donchian channel system, and go from there. I am learning how to use NinjaTrader, and have played around with R somewhat. There is so much information out there, that it is often overwhelming. My question is what resources would you recommend, i.e. books, websites, etc., that would be good for someone who is just starting. Thanks.
     
    Mattem and cjbuckley4 like this.
  2. I don't want to discourage you but think of how many thousands of traders have used these indicators to try to make money trend-following and where they have ended up. As someone point out in another thread, backtesting and analysis with these indicators gives you a false sense of reliability. The samples are very small compared to what you would need and there is no statistical rigor in any of this but just hopes and dreams. I think the greatest danger though is the opportunity cost from wasting valuable time. Good luck to you anyway.:)
     
  3. I would assume that any edge for basic trend following systems has been essentially arbitraged out of the market. It's more of a case of being interested in systematic trading, and needing to start somewhere. Thus, anything that can reduce the learning curve is to my advantage. I don't expect to build a perfect system right off the bat.
     
  4. I have built hundreds (if not thousands) of trading systems, some of them profitable, and concur that simply trying to build a trading system using indicator-based rules rarely (if ever) works anymore. I should also point out that many traders start out down this path, but eventually give up and instead learn how to read price action. I have done both, and, in my opinion, the latter is much easier.

    If, after these comments, you are still interested in building trading systems, it can be profitable, and I suggest to you the following:

    - At first, forget about learning technical indicators. Instead, learn basic statistics and computational statistics. Then learn how to model and analyze time series data. Also learn about digital signal processing (filters, data transforms, etc.). Knowing this will go a lot farther than simply coding up some technical indicators (which are based on these concepts anyway). This will take a while. I received my PhD doing things like this, and am still learning new and useful things.
    - Then, get the book "New Trading Systems & Methods" by Perry Kaufman. It is one of the most comprehensive books about trading systems.
    - Then, get the book "Evaluation and Optimization of Trading Strategies" by Robert Pardo, and learn about backtesting. It is one of the best. And note that you can indeed get reasonable statistics via backtesting. The issue that you have to be aware of is that prior market data is never a perfect representation of future data.
    - As far as resources on the web, many of them are useless. I have started compiling statistics on a number of trading systems at www.deepalgorithms.com that you might find interesting/useful (for free), but it might be a little while before I have many systems logged. I am still setting things up.
    - Finally, get a subscription to Technical Analysis of STOCKS & COMMODITIES. It publishes the latest ideas in technical analysis.
     
    Last edited by a moderator: Dec 16, 2014
    Butterfly, IAS_LLC and Baron like this.
  5. Thanks, that it some great information. I will look up those books. I wanted to use Ninjatrader as sort of "training wheels" for backtesting. I am not good enough in R or any other program to do backtesting as of yet. I do agree with the need to learn statistics as well.
     
  6. Your site has a Kick Ass Layout and Visual Appeal. Really good job on the design!

    Here is some stuff that works if you know how to use it: http://www.trade-ideas.com/trade-ideas-university/alert-definitions/
     
  7. I could not agree more, price action trading is the closest thing to the holy grail and delivers far more superior results than all the usual indicators combined.
     
  8. Could not agree more!
     
  9. Thanks for sharing. I like your site. Also, adding your recommended books to my reading list.
     
  10. Thanks for all the responses. There is a lot more info for me to review now!
     
    #10     May 1, 2014