I did not analyze anything. I just backtested EA as ex4 file with the settings/timeframe from post #4 (txt file). That's all. I did not optimize the settings.

If EA is profitable in 2006 only but not in 2005, not 2004 and so on - so we can not count it. It is not profitable in this case: market may be changed tomorrow to 2005 year market for example.

I used the settings in text files. But i think it should be different settings and some instruction about "do not trading during the news time" and so on.

No one EA can be profitable "forever".
I think this EA may be very profitable for particular market condition. In particular time may be ... i don't know ...