Monte carlo simulation

[Volker Butzlaff]

@goodgoing,

hmm...who don't get the message?

I've never said that I test stress test random systems. I test real systems (and the real generated system report results) and the MCS stress test shows their inherent possible behaviour under equal market conditions. Benefit of this is the knowledge, that profits of a system may be far lower than in the unique historical test and draw downs far higher...

The Monte Carlo Simulation testing technique generally uses (pseudo-)random calculations, but with real life models! That's not only in trading system development, that is a well known technique in all sectors, where you don't have enough real or historical data (or it's to expansive to simulate these scenarios in reality). Specially in trading system development you have often only 10, 20 years of proven quality data. With MCS you can simulate historical periods of 100.000 years and more...that's the point!

The second benefit is targeted to (possibly) changing market conditions. With additional generated test data you can do far more system tests as only with your unique historical data set. So you get a brider understanding of the robustness of your trading system setup. Clearly - not a 100% solution. But there's no such solution out. The possiblities are infinite.

Only my two cents...

bye,
zentrader

 

[black diamond]

Quote from Volker Butzlaff:

@intradaybill,

I use the MCS method in the trading system development process as shown below:
http://www.zentrader.de/mcsprocess_e.pdf

For me there are currently two valid benefits using MCS:
a) to generate alternative test data (data preparation phase)
b) to stress test the system/system report results (system simulation phase)

But you have to use the methods that are convenient to your development style ... and perhaps there's also a benefit in using it directly in the trading system setup as you explained before...
...i don't know it - but I'm always interested to learn something!

I think he is talking about a 3rd use of mc - a test for signficance using reshuffled data instead of a theoretical distribution.

There is also a 4th use I am aware of, to value path dependent derivatives like mortgage backed securities.