Trading the same system with different variables

Discussion in 'Strategy Building' started by jrkob, May 18, 2004.

  1. jrkob


    Let’s assume a given trading system called S.
    S has, for simplicity’s sake, 2 degrees of freedom, x and y.
    S(x,y) is therefore the trading system S using the couple of variables x and y.

    We optimize our system, back test, forward test, paper trade, etc etc… and we find that for instance S(3,5) and S(5,3) have both very acceptable results.

    Would that make sense to trade BOTH at the same time ?

    The reason why I’m asking is because in my trading system, I find that trading both would present a smoother equity curve, and lower historical drawdown. Of course, the return on equity, or yield, of trading both is lower than trading the best one.

    Anybody found any particular issue in doing so ?
  2. mmillar


    You're trading the same system with double the contracts.

    What's the difference between trading...

    1 contract of S(3,5) AND 1 contract of S(5,3)

    2 contracts of S(3,5) OR 2 contracts of S(5,3)


    Just looks like twice the risk, twice the reward to me.
  3. jrk

    It looks like you've answered your own question.

    For any 2 systems:

    - Return(s1+s2) cannot be better than max(return(s1,s2)).
    - D'down(s1+s2) cannot be worse than max(ddown(s1,s2)).

    So you're trading overall return for smoothness of equity curve. Whether or not this is a good/sensible depends on your acc size & risk tolerance.

    I'm assuming that you're trading same number on ctrcts in both scenarios.



    As stated the previous post, by trading both, you can only do better (in terms of smoothing equity) than you would just using the worst of the two. Trading a range of parameters which give similar results is a nice way to smooth your return. Now if you can duplicate the results for different stocks/currencies etc, for different parameters, I would say you would have some nice diversity!

  5. 2 cents: imo I'm wondering you're trading merely one system consisting of the same two variables (rather than different variables) with two different settings.

    Possibly, the two settings could be interpreted as an equivalence of a scale-up position size for multiple (in this case, two) contracts purpose. :confused:
  6. abogdan


    My 2 cents: Try your system (with the same set of constants) for two different stocks. If results are different then you are safe to assume that two different sets of settings applied to the same stock would produce similar behavior as the system would produced being applied to the same stock but with the different settings. If that is the case, then your assumption is correct; your equity curve should be smoother! Its called DIVERSIFICATION

    A long time ago I did similar analisys and managed to prove to myself that there is virtually no difference between diversifying your portfolio or diversifying your multiple strategies applied to the same stock.