molanisfx wrote:
If you have two variables, let's say tp (take profit) and sl (stop loss) you look for the best combination that give you the higher profit. Thus you select the darker area of the graph. The problem is that many people cross the line between optimization and over fitting (over optimization) - changing the setup so much to make the highest profit in testing but they forget about forward testing.
When building statistical models, one could use 2/3 of the historical data for optimization and 1/3 for testing after optimization. The issue here is that the latest market data has valuable information.
This is a nice article that talks a bit about common errors:
http://championship.mql5.com/2010/en/news/23Hi. Thanks for that reply. AM not new to trading generally (normally do manual trading) but very new to automating trades and will be learning more till i get better.
Have posted my final ea. Its just the simple rules i use in my daily chart. And the result i have here compare to that i see on mt5 forum is intimidating. People making profit of over 800k. Anyway i will still like to to participate at least for the fun of it - noting to lose.
Would have loved to use my other system which is on smaller time frame (it will enter more trades) but it requires 'pending order' and is not present here. Pls are you going to include it in the final version.
Note: My a/c in the ATC2010 still not withing the required a/c range. Am i to do anything about it or what.