Backtesting Metrics

Discussion in 'Automated Trading' started by DustyFoot, Dec 31, 2010.

  1. DustyFoot


    Hi Everyone. I have been working on developing several ATS and I was wondering if any of you could let me know what kind of metrics you use for your backtesting results before you decide to execute a strategy. Things such as Drawdown VS profit, profit factor , number of executions and how far back you look in order to consider your strategy robust enough. I would imagine this has been asked and answered but I was unable to find a thread. Any insight would be appreciated.
  2. nLepwa


    You only need to look at two things:

    -Rolling profit factor
    -Nb of trades

    The rest is noise.

    For an intra-day strategy I would look at 15+ years of data.
    For daily, weekly, monthly 25+ years is good.

  3. ronblack


    Why are the rest noise and profit factor isn't? All metrics are related in a way or another.

    Why 10+ years for intraday?

    Why 25+ years for daily?

    I think you should offer some justification for these bold statements of yours. You may be right but I have seen more people being wrong than right in this business. Bold statements without justification mean absolutely nothing to serious traders.
  4. Six months of data. Profits greater than 2.5X losses. Wins at least one time out of three. Results consistesnt for twelve rolling six month periods. Scalable.
  5. My favorite source is "Design, Testing, and Optimization of Trading Systems" by Robert Pardo. A bit academic in presentation, but full of ideas to help you move forward and to avoid fooling yourself.
  6. ronblack


    People who trade don't have time to write books about specifics of trading. As a result, it is axiomatic that books will not offer you any insight or edge. So stop fooling yourself. By the way, why don't you tell us something you learned from Pardo's book that can help us make money or trade more profitably?
  7. The op was addressing the issue of developing. That is sufficient justification for my recommendation.

    The rest of your post is what my friends in the military refer to as a "target rich environment." I'll give you a pass, as I'm off to my trading platform for a bit.
  8. ronblack


    I suppose the next time you are cornered you will talk about friends in other planets or universes or dimensions?

    Are you trying to scare people away from these boards?
  9. fjpenney


    You shouldn't expect a consensus answer on the important metrics as we all have different views. In my opinion, you can't consider return without considering risk and risk isn't volatility. For risk I look at the Calmar Ratio (CAGR / maximum drawdown) and Ulcer Index (as developed by Peter Martin).
  10. My trade filled, so I'm back.

    Cornered. Oh, I can't wait to hear an explanation for that. Do tell!

    I have no idea why saying I have friends in the military would scare anyone away from anything. My friends in the military primarily come from two of my areas of interest. I am a private pilot and make many acquaintances who are military. Also among air traffic controllers, commercial pilots and other private pilots. Cleveland Center is a wonderful place to visit. Not so scary yet.

    I am also a Concealed Carry Weapons instructor in the State of Ohio. One needs a permit from the state in order to carry concealed and must take a class in order to obtain one. I've met many in the military in that venue as they want to carry when out of uniform or after military service. That may be scary to some who do not believe in self reliance and self defense.
    #10     Dec 31, 2010