Fully automated futures trading

Discussion in 'Journals' started by globalarbtrader, Feb 11, 2015.

  1. You're right they are missing. I did the selection manually by copying and pasting off the IB website, and as I was trying to ignore single stock futures I may have missed some (in fact it looks like I've missed loads in SGX). I've got the XIN0I but not XINA50, I already have the Nikkei in Osaka, but I don't have the Nifty at all.

    I'm also missing (though haven't checked if these are viable yet): AJ, AU, CY, INR, IU, KJ, KU, SGB, SND, TD, TWD, TWN, UC, UJ, UY.

    GAT
     
    #2691     May 7, 2021
  2. Well I'm trading futures so there is already leverage. But I think what you are referring to is the fact I run my portfolio risk target 2.5x higher on each instrument to account for portfolio diversification effects.

    GAT
     
    #2692     May 7, 2021
    blink18 likes this.
  3. I'm looking at "-0.032/0.032" and the mid point is 0, therefore it can't be +/-10% (or any other percentage for that matter). What method do you use here?
     
    #2693     May 8, 2021
  4. It's based on the average position (with a forecast of 10).

    GAT
     
    #2694     May 8, 2021
    test_user likes this.
  5. I see this is a different method than in the book. I looked into your code and found it is "forecast" buffer_method (alternative to "position"). Do you use 10% buffer size as well?
     
    #2695     May 8, 2021
  6. Kernfusion

    Kernfusion

    Hi Rob, great stuff, thanks!
    I noticed that JPY-denominated Nikkei 225 with multiplier 100 from GLOBEX has high cost 0.33795084:

    upload_2021-5-8_15-51-54.png

    So you choose the other one with multiplier 500 (NIY) in your final list.
    I actually trade the OSE.JPN version of the x100 Nikkei (N225M), which you have in your list but apparently don't have data subscription for it. As already mentioned, that data is very cheap and that contract is pretty-liquid and 5 times smaller, so probably makes sense to trade it instead..

    Also, I guess you've decided that these new 70 contracts provide enough diversification (and work for adding them :) ) that you'll not going to bother with trading ICE contracts without the real-time data using IB-algo orders, at least for now ?
     
    #2696     May 8, 2021
  7.  
    Last edited: May 9, 2021
    #2697     May 9, 2021
  8. No. The 10% method is a simplification. I've had enough complaints about the book being too complicated without making it worse. But it makes no theoretical sense to have a different sized buffer if you have a max posiion on, or a zero position.

    GAT
     
    #2698     May 10, 2021
  9. Yes, I've had so much helpful feedback (mainly from @HobbyTrading) it makes sense to rerun the exercise now I've added the subscriptions for OSE and SGX, and made a few other tweaks.

    No (ICE) it doesn't make sense for now. But it's still on the table in the future.

    GAT
     
    #2699     May 10, 2021
    Kernfusion likes this.
  10. Ooops, sorry for creating more work. *frantically searching for the "blushing" emoji*
     
    #2700     May 10, 2021
    Kernfusion likes this.