Frosty's Strategy Analysis

Discussion in 'Strategy Building' started by frostengine, Dec 5, 2007.

  1. In this thread I will provide performance charts for strategies I am currently working on. The goal of the thread will be to get feedback concerning these strategies. Which should hep me decide which strategies have the best chance of succeeding going forward. I am welcome to all "constructive" comments regarding the likelihood of success based on these performance charts.

    All performance charts will be based on strategies trading the S&P mini (ES). The strategies are built using time bars that are sub 5 minutes. All results assume the strategy trades only 1 ES contract at a time. The historical data goes from: 07/31/2003-12/03/2007

    Here is some statistics about the first strategy "strategy1" that we will discuss.

    Number of Trades: 2438
    Win Percentage: 43%
    Profit Factor: 1.2
    Max Draw Down: -$4699
    Total PL: $32,847 (per contract)

    The first Chart I am providing is the PL chart.
  2. This chart details the profit and loss for each individual trade.
  3. This chart details a running total of the drawdown.
  4. This chart details the max draw down experienced for each individual trade.

    This is the last chart for this strategy I plan to post for now. Hopefully this turns out to be a meaningful discussion.

    Thank you for your time
  5. AaronCapps

    AaronCapps Global Futures

    Are you calculating any slippage and commission into your results?
  6. Yes the current commission rate charged by IB for ES futures and 1 tick slippage is being applied to all calculations. Some may say 1 tick is not enough slippage, but for the ES I find you seldom experience more than 1 tick of slippage during business hours. If someone has some input concerning slippage they routinely experience with ES I would be happy to hear it and adjust my calculations accordingly.
  7. I have a question for those of you who looked at the PL chart. If you notice the fist 1/4 of trades or so was basically only about break even. It wasn't until the next 3/4 that it started to really make money.

    Would this be a major concern for someone trading this strategy? One thing to note is that the strategy appears to be picking up steam in this current market.

    Any fundamental shifts for the ES between 2003-2004 to account for poor performance early on?
  8. i would only be concerned if the system would accumulate drawdown during that flat period of yours.

    otherwise the drawdown is a bit high for my personal taste but small bits of data you presented us looks ok to me.
  9. 1) The flat period would concern me unless I knew what was causing it. If you do not know, then some mechanism to switch off the system or otherwise trade the equity curve would be comforting.

    2) I'm not sure the system is picking up steam as such. It is just that volatility has increased so trading a 1 lot with a positive expectancy will return more $ now than it did in quieter times. Remember daily ATRs now are 2-3 times higher than early this year.

    3) By 1 tick slippage, do you mean 1 tick each leg of the transaction? What type of orders are you using to enter the market? If you are using market or stop orders then 1 tick is too conservative in my view. Unless routinely trading over economic figures slippage is rarely an issue until trading large size. If you are using limits then it's a fair assumption - you may not get filled even if the historic data shows your price trading, unless the market taded through your limit.

    4) Looks like you are netting approx 1 (and a bit) ticks per trade after slippage. This is ok but not ideal - with such fine margins even small errors (missing trades for example - power outages, etc) can make all the difference.

    Overall looks like you are heading in the right direction. 2.5k trades with a positive expectancy suggests you have uncovered some sort of unrefined edge. But for me the margins would need to be higher to action this type of strategy, along with some understanding of what was happening in 2003.
  10. The results look good.

    One thing i noticed is that you seem to using large stops (or no stops), which seems strange for a system that trades off sub five minute bars??
    #10     Dec 5, 2007