Frosty's Strategy Analysis

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

  1. the profit factor is too smal..This system does not give you a profit in future. You must remake it. And set the cost about 0.05%ore 0.1%.
     
    #11     Dec 6, 2007
  2. RhinoTrader,

    I am using 1 tick slippage per trade, not per side. And its using market orders to enter. Since the spread on the ES generally stays pretty tight at just 1 tick with large bid and asks on both sides I figured this should be sufficient. Perhaps I should throw in another tick of slippage in future calculations just to be safe....

    Also good point about it netting just over 1 tick per trade. I had not thought about it that way. You are right those margins are too small.




    Businessman,


    The strategy is still very "raw" at the moment. Right now the strategy is a "stop and reverse" strategy. No stops, no alternative exits. Simply reverse on opposite signal. And it also exits when US equities markets close.
     
    #12     Dec 6, 2007
  3.  
    #13     Dec 6, 2007
  4. As would be expected after Rhinotrader pointed out the strategy was only yielding a little over 1 tick per trade, adding an additional tick of slippage was detrimental. The strategy only produced a little over 2k of profit in this scenario. Attached is the PL chart showing how things played out.

    I know the strategy is far from ready for live trading yet, but the fact that it can overcome full commission, plus 2 ticks of slippage over almost 2500 trades and still be slightly profitable suggests there is "some" edge in there.

    Now I need to decide which path to take to "refine" this edge. The first thought that pops in my head is why not simply add a hard stop. However, All my past experiences have shown me that stops almost NEVER improve a system. BUT I will give it a shot and post my findings.
     
    #14     Dec 6, 2007
  5. In my next 2 messages I will show some draw down charts. The first chart shows draw downs only for WINNING trades. Like this I can see how much heat the trade took even though it eventually made money.

    The second chart will show the draw downs for only the LOSING trades. Like this perhaps we can draw some conclusion as to where a good stop point may be.
     
    #15     Dec 6, 2007
  6. Looking at the 2 charts, a good cutoff point I see would be $200. Doesn't appear that many winning trades go past -$200 and still end up positive. However, looking at the losing trade draw downs we see that a lot of them go past -$200. So the first stop we will try is hard stop right at $200 draw down.
     
    #16     Dec 6, 2007
  7. BTW all charts i'm showing again are using only 1 tick of slippage for easier comparison with the first posted charts.

    After putting in a hard stop at $200, the results have got worse.

    Total PL: $22,104
    PF: 1.12
    Win%; 41.6
    Max DD: $5136

    Anyone has any theories on why the hard stop affected it this way after looking at the 2 draw down charts?

    I'm guessing that the winning trades that had more than a $200 draw down were some "big winners" after going positive. Perhaps at those times volatility was much higher causing it to drop below $200 but when it went positive it went way positive.

    Any suggestions for other hard stop points? instead of trying to "guess" a good spot for a hard stop I will now have the strategy optimize itself attempting stops from -$10 to -$2000 in increments of $10 and see if any of the hard stops yields better results.
     
    #17     Dec 6, 2007
  8. As I expected and as was confirmed by the optimizer, no hard stops increased performance at all.

    Does anyone have other ideas on how to refine this strategy?

    The next steps I could try would be to find filters for entries.... also to try and find other ways of getting out of a trade early besides a monetary stop.
     
    #18     Dec 6, 2007
  9. Here are the performance charts of 2 more strategies. The first one "strategy2" uses the same indicators as the original strategy of this thread. However it has been tweaked for better results. The second strategy chart I'm about to post "strategy3", uses different indicators than the first 2 but the concept behind the strategy is the same. Just applied different indicators to the same philosophy. Both of these strategies are better than the original. To better compare with the original all of the calculations include same slippage/commissions as before.

    Strategy 2:
    Total PL: 35,347 per contract
    Trades: 1899
    Win%: 51
    PF: 1.27
    Max DD: 4060

    Strategy 3:
    Total PL: 19,799 per contract
    Trades: 672
    Win%: 51
    PF: 1.34
    Max DD: 3140
     
    #19     Dec 7, 2007
  10. Here is the chart for strategy 3. All comments and questions are welcome.
     
    #20     Dec 7, 2007