Robot Trading on futures - ES et al

Discussion in 'Journals' started by iuykcif, Sep 3, 2009.

  1. the viability of the strategy depends on the entry and exit. remeber OP stated somewhere the entry is random, and I doubt it will be workable in that way. I like to see I am wrong, but so far I am still convinced it is no different from normal averaging down. the 2 contracts limit the loss but at the same time limit the profits. so regarding profit and risk, it does not increase the ratio than in normal averaging down procedure.
     
    #61     Sep 10, 2009
  2. iuykcif

    iuykcif

    Thanks ScoobyStoo,

    The averaging that may appear is not sought in any way by the inner logic of the algorithm. It's a mere byproduct. It simply stems from the fact that the double edging / scalping action has to let increase the position on one side to continue the scalp on the other side. Also it does not occur necessarily, not even in a rally. It may or it may not, it depends if there are intermediate oscillations large enough which lead to a profit.


    > in a limit locked market or generally in liquidity vacuums

    can you elaborate a little bit more this part. Can you make specific examples (instruments, etc.) of what you are referring to ?


    Tom


     
    #62     Sep 10, 2009
  3. iuykcif

    iuykcif

    As I never leave anything unattempted, I am making an experiment forcing all orders to one port (that is, the same account).

    The result is "weird looking" and can be seen in the pictures below:

    [Contraints: Max Contracts: 6
    ScalpRange:1,2,5 ]


    <a target="_blank" href="http://www.datatime.eu/public/TradingDiary/ES-2009-06-11_CoupleHours.gif"> ES-2009-06-11_CoupleHours.gif </a>
    <a target="_blank" href="http://www.datatime.eu/public/TradingDiary/YM-2009-06-11_CoupleHours.gif"> YM-2009-06-11_CoupleHours.gif </a>
    <a target="_blank" href="http://www.datatime.eu/public/TradingDiary/NQ-2009-06-11_CoupleHours.gif"> NQ-2009-06-11_CoupleHours.gif </a>


    Looks like at first sight the procedure will not lose it's effectiveness.

    If this is really so, I need to study how to remove the "redundant" trades to improve the spread and commission expenses.

    Also, looks like if we could use on the other account an instrument with a "specular" price curve, we could be protected without options (?).


    Tom


     
    #63     Sep 11, 2009
  4. This is what I have thought that two accounts are only eye candy, no real sustance. what do you mean scalprange 5? does it mean in the case of short, whenever the price increases 5 points, you average down with another contract short?
     
    #64     Sep 11, 2009
  5. Congrats on a very interesting, weird and even potentially profitable thread.

    With regards to your question: Could you "force" the second account to be net long YM, when the first account is net short ES?. (In the context of routing all orders for same product into a single account).

    That could provide the "protection" you seek without using options.

    Do you "hedge" with market orders or limit orders?.
     
    #65     Sep 11, 2009
  6. iuykcif

    iuykcif

    Well perhaps from an operative point of view.

    Problem is that the logic has been though for 2 accounts and "then projected" onto 1.

    There is perhaps an audacious analogy with the math you do passing through the complex plane. You may return to real, but won't get the same result if you never leave the real line...

    The ScalpRange (or "ScalpSize" perhaps is a better name ?) is the min amplitude of oscillation captured by the scalper.

    smaller ScalpRange -> more Contracts -> more profit

     
    #66     Sep 11, 2009
  7. iuykcif

    iuykcif

    Thanks shikantaza,

    >"second account to be net long YM, when the first account is net short ES"

    hmmm, why do you have this intuition that it could be profitable?

    Are you thinking that, since the scalping process is symmetric, we can still profit from an opposite game and at the same time obtain a position approximately opposite to the other account ?

    Like:
    Acct1 N
    Acct2 -N

    Acct1 -N
    Acct2 N

    Can we think of a "specular" price curve to apply exactly the same process and be more confident we attain such a perfect hedge condition at any time ?

    However, if your intuition is right we need not use a different instrument (YM): we can still use ES for closer symmetry (it's another account). [Or still ES with another expiry date. ]

    Tom

    PS. No, I never use mkt orders (only in "manual mode" the user can send mkt orders if he wishes)

     
    #67     Sep 11, 2009
  8. You got it.

    Just "start" the second account with YM in the opposite direction of the ES account. Your "oscillations" seem to be in the "noise range" anyway. Who cares what the longer term direction is?. Right?. When a big move comes, you'll make a bundle in one account and lose an smaller bundle in the second account. If a catastrophe strikes, you'll be approximately flat.
     
    #68     Sep 11, 2009
  9. iuykcif

    iuykcif

    Hmmm, I am trying to visualize it mentally, but I still don't see it.

    In what you say I see a problem:

    - If you play an opposite strategy on the same curve shape
    it seems to me you are going to run an unprofitable strategy
    on one of the 2 instruments (intuitively can't be both profitable
    with a specular game on the same curve)

    It seems to me that to be "perfectly" hedged * at any given time *
    we do need instruments with "specular" price curves.

    So that both will be running profitably, but when one has a positive excess
    (because preparing for a reversal up), the other will have a negative excess
    (because loading for a reversal down).

    Or, perhaps I am missing to see some deeper meaning of your intuition ...

    What do you think ?

    Tom

     
    #69     Sep 12, 2009
  10. y=C-X (where C is a constant) is "specular" of X?. If not, please define "specular", preferibly with a graph, picture or drawing.
     
    #70     Sep 13, 2009