In the historic of Monte Carlo VAR methods, if the probability time number of observations is a fraction, how do you deal with that? For example, if the number of observations is n and the probability you want is 5%, and 5% times n is, say 1.2, do you round it down to 1, or jack it up to 2?
You could do an interpolation. Take the number 1 number take 80% of it then take the number 2 number and take 20% of it. that will give you the 1.2…I think. If i remember Level 1 Quant…
don’t think they’ll go there… but I would take the in-bewteen amount…
In other words to my post…it is 1/5th of the way between the two, so if you take the spread between the two numbers say 100 and then divide it by 5, so 20 the 1.2 number will be the 1 number say 45 + the 20 = 65…
just took CFAI online exam 2 and scored 63%…not great…but its better than the 43% i scored on my freebie…and just to tie that into this post…i will say that there were some questions on there about VAR
Quantitive or Qualitative type questions?
they were more qualitative…there was 1 that was very tricky…bcuz they gave diff #'s in the answer…so i’m plugging away doing diff calcs…nothing is matching…then i took a closer look…and by process of elimination…you had to figure out VAR based on relationships…(i.e. amounts and time horizons)
in the monte carlo var, schweser says that u use the model to get the expected return and std dev… and then you plug it in the analytical method… but one of the cfai end of book ques… uses monte carlo and applies historical var… any ideas?