Adventures in Automation

Discussion in 'Journals' started by fan27, Jan 21, 2019.

  1. fan27

    fan27

    I am going to revisit the hedging strategy. The original testing was a bit flawed as I was using one minute data. I need to use TICK data so I can determine the exact tick(s) where the portfolio open positions go into the red. The concept is to fade the open and losing positions while the open positions as a group are down. Should know something by Monday.
     
    #251     Jul 20, 2019
    They likes this.
  2. Where are you getting your tick data from? Also, why do you think tick data will help lose less money than 1 minute data? Why not just set your stops at the red?
     
    #252     Jul 20, 2019
  3. fan27

    fan27

    TICK data from IQFeed. First, let me state the problem I am trying to solve. My strategies are highly correlated and I already have a method for limiting the number of concurrent positions. However, my wins and losses are often grouped together resulting in an equity curve that is not as smooth as I would like it to be. My plan is track all open positions via a C# extension I wrote for AlgoTerminal. A singleton class keeps track of all of my open positions and their state and the class can be accessed in any strategy that is running. My strategies will still use 1 minute data as the source data but to test this in a simulation I need TICK data so the hedging strategy will know the moment the group of open positions go into the red. I do not want to set my stops at the red because I do not want to exit the position. Let's say I have the following positions:

    2 ES
    1 NQ
    2 YM

    Let's say at the moment the group of positions go into the red, the NQ and YM positions are in the red (ES is up). The hedging strategy would Short 1 NQ and 2 YM until the group of positions are in the green or until the target position is stopped out. The four positions that I had go bust just the other day would not have been a loss because while the positions were going down I would have been net flat. The obvious problem I need to over come is a situation where the group of positions go from green to red to green to red to green to red. etc. This would result in multiple Short trades where I could in theory get chopped up with very small losses. If this can be overcome, I think this approach makes sense.
     
    #253     Jul 20, 2019
  4. I think this is going to be painful for you but I hope I'm wrong...

    If your strategies are correlated, why not find uncorrelated strategies or trade different markets? That to me seems like a better hedge.
     
    #254     Jul 20, 2019
    fan27 likes this.
  5. fan27

    fan27

    It's not going to be painful. Either the research will bear fruit or it will not. Yes....expanding to different markets is on my to-do list.
     
    #255     Jul 20, 2019
    nooby_mcnoob likes this.
  6. d08

    d08

    That seems to be the right way to go about it. When trading indexes, I always choose the one's with least correlation. The main three US indexes are strongly correlated.
     
    #256     Jul 20, 2019
    fan27 likes this.
  7. fan27

    fan27

    Using the TICK data to test my hedging strategy is a bust. The issue is that the equity curve gets destroyed by the number of small losses. Moving on to other markets for diversification as others have suggested.
     
    #257     Jul 21, 2019
    nooby_mcnoob likes this.
  8. fan27

    fan27

    Exit 1 ES at 3018.50
    +$1595.9
     
    #258     Jul 24, 2019
    nooby_mcnoob likes this.
  9. fan27

    fan27

    Launching a new strategy portfolio on Monday with 135 strategies. I have made some enhancements to the AlgoTerminal code and added enhanced reporting to my Slack account. Still long 1 YM contract.
     
    #259     Jul 28, 2019
    Snuskpelle and nooby_mcnoob like this.
  10. fan27

    fan27

    Exit 1 YM contract at 27245
    +$629.90

    Currently flat.
     
    #260     Jul 29, 2019
    nooby_mcnoob likes this.