how do you evaluate the expected profitability of a portfolio of automated trading sy

Discussion in 'Automated Trading' started by jjk2, Jan 16, 2008.

  1. jjk2


    how do you evaluate the expected profitability of a portfolio of automated trading systems that each works in different types of market mood (system 1 trend follower, system2 range on)
  2. pneuma


    expectancy = (number of winners * average profit) / (number of losses * average loss)
  3. If you have system1, system2, and system3.

    The expected return for system1 in one year is P1 ,average DD is D1 ;
    The expected return for system2 in one year is P2,average DD is D2;
    The expected return for system3 in one year is P3,average DD is D3;

    Assuming the 3 systems are not related to each other, what is the percentage of money, w1, w2, w3 (where w1+w2+w3 =1) assigned to each system to get the maximum return/DD?
  4. This is a simple problem that can be addressed objectively. What you're looking at is 3 risk weights. You need to find an optimal combination. The simplest approach would be to maximize Sharpe ratio. This is how you can do it:

    1) take monthly results of the systems
    2) run through all combinations of w1, w2 ant w3, from 1 to 100. An example of a single combination: w1=2; w2=16; w3=52. With three systems this would be 100*100*100=1'000'000 combinations
    3) for every combination, multiply system's monthly results by its weight (w) and calculate monthly Sharpe ratio of all systems combined
    4) eventually wou will find a combination of risk weights where the monthly Sharpe ratio is max.
    5) normalize risk weights to percentage risked. For instance, if you find the best combination is 2;4;1, you may choose to risk 0.5%, 1.0% and 0.25% for the systems.

    What I've just wrote is just a conclusion of "System development with Acrary" thread. If you want more details, find and read it.

    Hope that helps.