beathespread

Yes, I agree

a solution could be a bunch of indicators from the fractals class as it's written 

but, sure those algorythm requires extensive use of data mining and a lot's of work of tuning, optimizing parameters.  In a practical way, as far as I see, big computers w/ big CPU and high memory are mandatory. 

To perform this task, this should be made in C or in C++ or C#. In my humble opinion, MQ4L is now obsolete, or it must be connected to a database and a mathematical software like mathlab

The question I ask myself

How could you know that the whole optimizing efforts were not in vain?

Below those posts are my main ideas on the topic

http://beathespread.com/blog/view/16854/optimization-of-trading-strategy-is-there-something-wrong-part-2#comment_7167

It looks like some strategies produce much more positive randomely generated results than others.

This can be used as some kind of robusteness of the strategy. Comparing spinal implant with asctrend both using the logistic kernel I found out that the spinal implant is producing less randomely generated positive equity flows than the Asctrend.

The optimization could be made w/ the genetic algorythm

I just upload it

Please, Do not take very seriously this post but I think Douglas Adams had much deeper insights than we may think.  Especially the idea that we humans we are designed to find the right question, not to find the answer.

The answer will be given by machines.

I checked what is available on FF regarding the use of Monte Carlo simulation. Here I found spread sheets for Monte Carlo simulation.
However I think not every system can be modeled using those spreadsheets. 

In those spreadsheets it is about modeling system expectancy. So it would be OK if your system has fixed stops and take profit levels so you can calculate exactly the win/loss. By doing that you can model it. 

However if you use stop and reverse signals and no stop loss, or if you use a time filter or trailing stop. I think that in those case the system expectancy is not modeled properly.

Yes, the Monte Carlo is one of the new topic.

I would like to have a look on Levenmark algorythm too

I was experimenting with excel sheets of Monte Carlo simulation but they were doing a diferent thing. Based on the performance of the system: % of profit trades and profit factor it simulates the hypothetic performance of the system. But this is something else and I wonder of it possible application for the evaluation of trading systems.

However the tool is interesting to see how the changes in the % of profit trades and profit factor affect the possible equity curves.

There are some paid excel addons doing selection without replacement monte carlo simulations. I even found a free video tutorials but it looks quite tricky for me and requires some coding for excel to get something usefull.