Automated GBPUSD Long Term Strategy (Tradestation Performance Report)

Discussion in 'Automated Trading' started by rttrader11, Jun 10, 2012.

  1. Tradestation allows for back testing from October 2002 and I have attached a weekly chart showing this time period.

    Attached is the performance report with the following commissions factored in,
    $30 Round Turn Spread
    $20 Round Turn Slippage
     
    • et.zip
      File size:
      437.1 KB
      Views:
      188
  2. Please do not post zip files. Post jpg files only. Only risk takers will download your zip file and click in its contents.

    Anyway, what type of timeframe are you talking about, how many trades did you get and what was the expectancy and win rate?
     
  3. gmst

    gmst

    Not sure what your motivation is for putting this system out there. Anyways I will write my opinion for your benefit and for benefit of others.

    It is a bad system. Might be tradeable in combination with other systems/markets (say you are trading 10+ systems/markets together, in that case you can trade it). As a standalone system, it is just too bad.

    1. Total Profits = 66k, biggest trade profit = 47k, so without this outlier trade, profits = 19k, total number of trades without outlier trade = 52, so average trade profit = 365$, which is 365 pips per trade (since trading vehicle is GBPUSD and initial equity is 10k USD). Average profits per year without outlier trade = 19k/10 = 1900 USD, which is a 19% return on a 10k USD investment.

    2. Total number of trades = 52, years 10, so 5 trades per year. Very unrealistic even for a swing trader to just take 5 trades in a year (unless and until he is trading multiple other systems).

    3. max intraday DD = 15k USD, so account size is less than starting capital, which means no leverage can be used on this account safely as risk of ruin exists. Looking at historical DD, this system has to be traded on a minimum of 20k investment, thus per year return around 19%/2 = 9.5% per year. Bad for a trading system.

    4. Longest flat period = 327 days, so system doesn't trade for 1 year at all. Again Unrealistic.

    5. Looking at performance graph, it sucks. 2003, 2009, 2011 and 2012 to date are losing years. Makes the system untradeable.

    If anyone has a different opinion about the tradeability of this system, I am all ears.
     
  4. Hi Intradaybill,

    I thought this forum was full of Risk Takers?
    I can see there has been 17 downloads and no complaints but I will try and attach an xls file for you.
    This strategy uses intraday and daily time frames to enter the market and then looks to hold the position only Exiting when the day trend suggests too.


    Hi gmst,

    Thanks for letting me know it’s a bad strategy, I’ll just throw it in the bin.
    Maybe I will wait though, Your opinion seems to be flawed, Max DD is $1.5k not 15k. So the strategy can be used with a 10k account.
    Can you post a strategy performance report that is good,
    You seem to know so much about it I would be very grateful if you could show us all how your strategy traded this market over the same time frame Or any market over this time frame.
    Ps/ Not sure what your motivation is either.
     
    • fts1.xls
      File size:
      82.5 KB
      Views:
      106
  5. gmst

    gmst

    I will post one of my system reports when I get some more time.

    For the time being, a quick point for you. Your DD is 15k on an open trade basis, so you will get a margin call and in this electronic world, broker would liquidate you immediately if you have < 15k in your account. On a closed trade to closed trade basis, yes your DD is 1.5k but that is irrelevant. I hope you understand what I am saying here. (Hint: compare left and right side of DD report on the first page of TS report).
     
  6. gpiu

    gpiu

    Hi rttrader,

    IMHO optimization reports on their own are of little to no value. It is always possible to generate a better glowing report that has no statistical significance.

    Your algo generates so few trades that it should be straight forward to perform a WFO. This should shed more light on your algo's ability to generalize.

    As an example, I posted a screen shot of an ongoing analysis. It depicts an explicit walk forward result of trading 10 ES contracts ~15K times in 2001 with $2 commisions and 1 tick slippage on each side.

    To put it in context, none of my strategy reports show a PF > 1.7 over the same time period. What is more relevant, how robust is your system to the market's non stationary nature.
     
  7. Hi gpiu,

    Your DD looks massive over 10k?

    My strategy handles sideways markets by not trading them.


    rttrader11