To whom may be so kind and answer my question. I am currently exploring the world of MC-simulations and matlab, which is very interesting. In this example/question I use the share price which I have logaritmically counted. And I coung with copula functions the distribution and get the answers [u,v]These two are then transfered with Normsinvu=Normsinv(u) and the same for vNormsinvv=Normsinv(v)Then I calculate the price.. Price1 = 0*exp(Normsinvu/100)Price2 = 0*exp(Normsinvv/100)Correct so far?? Well then I receive the price for the two shares which have entered the portfolio with one share each.. Portfolio = [Price1+Price2]From this I choose the maximum 5% for my 95% confidential level. At that point is my VaR value. Correct?Is this correct or should I revalue the portfolio at anytime? I count for the timeperiod [t,t+1]That data is then backtested against the realshare for 1000 tradingdays and I count how many times in % those are over my VaR. Which should be close to confidential level. The closer the better.. Is this correct way? Thank you in advance Thomas