The Problems with Monte Carlo Simulation

[Warren Buffet II]

I found this interesting article about the problems with the Monte Carlo Simulation.

http://www.fpanet.org/journal/articles/2001_Issues/jfp1101-art12.cfm?renderforprint=1

I like some parts of this article:

"The problem is that the typical assumption set used in Monte Carlo simulation assumes normal distributions and correlation coefficients of zero, neither of which are typical in the world of financial markets"

"Unfortunately, the parents turn into grandparents with foggy memories, and the children into parents who don’t remember the toys that broke. So several generations later, the old toys are back under the tree waiting to break again. ... Today, history is repeating itself with nifty spreadsheet add-ons that do all of the heavy lifting for Monte Carlo simulation and with the converts thrilled with their new toy"

"The problem is the confusion of risk with uncertainty. Risk assumes knowledge of the distribution of future outcomes (i.e., the input to the Monte Carlo simulation). Uncertainty or ambiguity describes a world (our world) in which the shape and location of the distribution is open to question. Contrary to academic orthodoxy, the distribution of U.S. stock market returns is far from normal."

Very interesting

 

[tech/a]

True Montecarlo simulation takes into account every known variable of an event.I certaintly agree that this isnt done in financial markets as Im un aware that every variable can be identified or has been for that matter.

What I do know is that so far all parameters in the methods I trade are within those parameters give with the known variables in the Montecarlo simulations run on the methods.

Until trading falls outside those parameters Im happy to accept possibly flawed assumptions.

 

[doctorj]

An interesting article, but no offence when intended to the author when I say - duh. Any system that involves simulating a dynamic real world system on a finite set of assumptions is going to be flawed.

My expertise is clearly not Computer Science, but linear increases to the complexity of such simulations results in a non linear increase test times but at the same time, greater detail concerning the nature and distribution of outcomes is highly desireable.

In the mean time, MonteCarlo and others provides a means of testing where equations are either unknown, not solveable or involve many variables. I've only read about half the article, but does it suggest any suitable alternatives?

 

[Warren Buffet II]

True Montecarlo simulation takes into account every known variable of an event.I certaintly agree that this isnt done in financial markets as Im un aware that every variable can be identified or has been for that matter.

What I do know is that so far all parameters in the methods I trade are within those parameters give with the known variables in the Montecarlo simulations run on the methods.

Until trading falls outside those parameters Im happy to accept possibly flawed assumptions.

Hi Tech/A,

I have actually read some of your stuff about your trading strategic and what I am saying (Myabe people got it in the wrong way) is that you have built your system around this simulation and it has worked for you. I find that interesting because you have been able to identify some parameters which are not known by many or at least have not been interpreted in the way you did.

As I said before, I have read some of your stuff but it is only fragments.

Regards,