<t>Dear all,Thanks for the many and varied inputs. Another thread (sub-thread? threadlet?) I think would be interesting to explore is the area of simulation using <b>quasi-random</b> as opposed to (pseudo-) random numbers. Of course, this has been covered in other threads, but for completeness it mi...

<t>Some say Mulligan's is the best pint in Dublin. I've not tried it. In $n$ years of being a legal drinker, I've never been able to get inside Kehoe's, so crowded is it. The Buttery? Dublin's surliest barmen and dungeon ambience. I hated the place. (Spoken as a Trinners alum, I might add.)Ron Black...

<t>My €0.02 (or IRP 0.0156):Belgian beers: Chimay Bleu is hard to beat, but Rouge is a somewhat dirty. (At least the bottles I got last weekend, but possibly those were unrepresentative.) Also, Orval is quite good, and comes in a wonderful glass, and Triple Westmalle hits the spot too.Budweiser: An ...

<t>Yes, the original problem still remains unsolved; all I meant was at least I understand what's happening now. (See, the original code behind the sheet, which I didn't write, had a fixed seed, so returned the same value every time. When my code didn't, I got confused, not knowing how Excel's RNG w...

<t>QuoteOriginally posted by: MikeCroweYou never really answered Halliron's question which is critically important here. I said 'every run' gave a different result, by which I meant that I got values x_1, x_2, x_3, etc.I agree with you --- it is concerning. Having read what Numerical Recipes had to ...

<t>S., T., thanks for your wise words. As for the stdev check, I've done that; the problem was definitely not the number of paths. It seems as if serial correlation of Excel's rands is the most likely explanation. Also, the draws are multivariate (62-variate, in fact) --- perhaps this also contribut...

<t>Thanks for the comments. The issue has been cleared up, insofar as the numbers I get are now stable when I rerun the macro. The problem was that the seed wasn't being set in my macro, and so it was using the system time as the seed for the generator. Hence, instability. (H --- every run gave a di...

<t>Thanks for your prompt reply. How would I go about replacing the built-in generator?As for the algorithm, what happens is roughly as follows:1) A vector of uniform random numbers is generated, then transformed according to a normal distribution (with given correlation)2) The transformed random nu...

<t>Hello All,I'm doing a very basic Monte Carlo (via an Excel macro) to implement the normal copula model. Now, the problem I have is that I can't seem to get a consistent result--- I'll run the simulation once, and get an answer x_1, then run it again and get an answer x_2 which isn't a million mil...

<t>My €0.02: Since you have the correlation for the 0-10% tranche, you can work out the breakeven spread on that tranche (via Monte Carlo or whatnot). Now, since the 0-10 tranche is just the sum of the 0-3, 3-7, and 7-10 tranches, the breakeven spread on 0-10 is just the appropriate weighted average...

<t>Hello all,I'm trying to generate a normally-distributed random variable by applying the inverse of the cumulative normal distribution to a uniform (0,1) rv. The approximation I'm using seems to be the standard one, of this form:where t = sqrt(ln(1/p^2)) and the c's and d's are ugly constants.My q...

<r>Hi vs83,Galiani's Master's Thesis is a good introduction, with discussion of various copulae, but perhaps more financial than purely mathematical. Find it here:<URL url="http://defaultrisk.com/pp_crdrv_41.htmLi%27s">http://defaultrisk.com/pp_crdrv_41.htmLi's</URL> paper is often quoted as the 'se...