Somewhat puzzle on porfolio test procedure and MC
Posted: Sat Dec 17, 2005 12:10 am
Recently I met a problem about sequence of testing porfolio,position sizing,and monte carlo simulation.
When running a trading strategy on a single security in a software,and accepting its performance,I used to test it on a porfolio to conform its performance robustness.
But this time,different securities results different profit,the weights of their contribution to the total profit are different,too.Running Monte Carlo simulation with these data would distort the performance(That's,the result of MC simulation shows mainly the effect of those with the most weight).
Should I normalize the position firstly,and then perform the monte carlo simulation?
When running a trading strategy on a single security in a software,and accepting its performance,I used to test it on a porfolio to conform its performance robustness.
But this time,different securities results different profit,the weights of their contribution to the total profit are different,too.Running Monte Carlo simulation with these data would distort the performance(That's,the result of MC simulation shows mainly the effect of those with the most weight).
Should I normalize the position firstly,and then perform the monte carlo simulation?