What are possible ways of extending & use "Artificial Intelligence Expert Advisor"?
I just wanted to become a little bit more familiar with neural networks in EAs so I have downloaded your EA to have a look at it. I've downloaded data from http://www.forextester.com/data/files/GBPUSD.zip
(and used period_converter to create GBPUSD H1 from M1) and changed two settings - lots and initial deposit (to reasonable values 0.3 and 5000 EUR).
I have backtested it on the period 2012.01.01 - 2013.02.08. What I can see is loss of 42% in that period (the same results for both models 'every tick' and 'open prices only').
How am I supposed to use this EA? There are four parameters (and SL as fifth one) so that would be terribly time consuming to backtest it (however, if 'open prices only' model is acceptable, it would take much less time).
Has this EA been used for real trading or is possible to use on real account?
I am not afraid of making programming changes to its code but I need to know a little bit more about some main ideas behind it.
I have spent some time on trying to add auto-optimization feature to the other Expert Advisor. So I know how to change values in the MQL code. I have decided in that other EA to create external application in C# to help me with backtesting auto-optimization feature.
I thought that auto-otpimization should be something like - you have two variables to optimize (here we have four variables so that's too much to manage it easily), you have their ranges (start, step, stop), you perform optimization regularly (for example every day at the same time), you change parameters after that optimization and so on.
How does idea connected with this expert advisor differ from what I have just described above?