Trading The Equity Curve & Beyond

About the Author Jeff Swanson

Jeff is the founder of System Trader Success - a website and mission to empowering the retail trader with the proper knowledge and tools to become a profitable trader the world of quantitative/automated trading.

follow me on:
  • Mark says:

    Jeff, you recommend to “test different moving averages and test between halting all trading or reducing contract/share size.”

    Is this not curve-fitting? If it is then is that particular moving average probably going to help or hurt me going forward?

    • IT could be. Anytime you pick a parameter it could be curve fitting or it could just be optimizing. What I often recommend when testing a range of values, a look-back period for a moving average for example, is to test the robustness of the parameter. That is, I would expect to see many different values that produce favorable results. However, if simply changing the look-back period a small amount drastically changes the results, I would be worried about curve fitting. If or when I find a region which appears stable, I will then take the mean or median value.

  • Ryan says:

    Jeff, have you taken a look at MSA software? It will allow for monte carlo testing using an equity curve.
    http://www.adaptrade.com/product.htm

    I did not find equity curve trading worked for me as the curve can take you out just as the best trades are setting up. Especially, when you have strategies where 20% of your trades make up 80% of your profits.

    Ryan
    http://www.daxgaptrading.wordpress.com

    • I have a copy of the MSA software. I really like it. That’s a good observation about where “traditional” rules for trading the equity curve may not work. In this type of situation you may find it interesting to test increasing your risk when under the equity curve average and reducing risk when it’s above.

  • Actually it’s a completely re-written article written in the last few days! The original article from 2012 did not have a equity curve solution. It only talked about the topic as a theory. There was no solution. So, once I found an EasyLanguage solution I re-wrote the article as the old article become completely obsolete. Thanks for the links to Kevin’s articles. I have yet to read them. It’s true that this technique may not work for all systems. I would never expect a single technique to work on all markets. It does not even need to work on most markets to be valuable. If it only works on less than 5% of strategies developed, it’s worth keeping. Remember, this is just another tool in your toolbox that you may find helpful. I would test this technique like I would test any other technique. How do you test that a trailing stop works or an open range breakout works? What steps are taken to boost the probability of success when incorporating those techniques in your trading? The same steps would apply here. There is nothing different. In all cases, your own system development process should be applied.

    • Thanks for the link! I’ve heard of this concept before and it’s really an interesting idea. In fact, you could use the Equity Curve Feedback tool to do something similar. For example, you could have a single strategy which trades both a mean reverting strategy and a trend following strategy. Using the functions within the Toolkit you could have it switch between these to types of trades based upon historical performance in real-time.

  • LB says:

    Hi Jeff,
    does it work also with limit and stop orders?

    If i am not wrong it works only with market orders….

  • >