My question is: what's the difference between running an EA in a productive environment compared to a simulated environment such as the strategy tester?
I'm thinking of aspects like
- quality of modeled environments (data quality for instance)
- availability (hardware, power, internet connection, …)
- predictability of the profitability of an algorithm
- etc. ...
Also I wanted to say, that I'm really surprised to find such a vivid and productive community here.
The MQL4-book is also a really great resource for beginners like me!
Productive vs. Simulated
The number one problem I've discovered is that when running a EA in a strategy tester- you will find issues of "Curve Fitting." What that means, is that the test shows good results for the past because the EA worked for the trend- but doesn't actually work period.
When making or using a good EA. Make sure it uses sound money management, takes sound risks, and above all else- uses sound discipline.
most time, or till now, there is no any uniform mothods to evaluate all such things.
you can defined yourself