|
PureBytes Links
Trading Reference Links
|
Mark,
that's an interesting article. The bespoken methodology by itself makes
a lot of sense.
But what makes me nervous is the fact that it works on the equity curve.
This equity curve could be derived from an (accidentally) curve-fitted
trading model. So doing Monte-Carlo over such an (potentially)
ill-derived dataset would not make much sense any more. - Garbage in ->
garbage out
I am also constantly thinking about applying Monte-Carlo-simulation on a
trading model which is based on a portfolio of markets. This is IMO very
hard because the various markets are or could be intercorrelated with
each other. So scrambling the equity curves of single markets and
deriving a portfolio value for that would just flatten the results.
Wouldn't it be much better to "scramble" the market data by itself.
Doing that n-times and calculating the cumulated performance statistics
would maybe make more sense.
The problem by itself is the "scrambling" of the market data, which is a
discussion of its own.
Maybe the same methodology applied by Mark to the equity-curve would
also make sense to the market data as well.
ANY COMMENTS ON THAT?
Another way for "scrambled" market data could be the methodology
described by Mandelbrot in "The (mis)Behavior of Markets". But
unfortunately there is no program code to work from given in its book.
HAS SOMEBODY MORE INFORMATION ON THAT METHODOLOGY?
Regards,
Christian
|