Why so many different opinions :)
Because humans always have different contents of their brain :))
How great our God ^_^
made wrong opinion to be true
Really that's ok. I wasn't trolling about competition between platforms, nor MetaTrader's coders nor others platform's coders judged, for sure every coders doing its best, and every traders finding its happiness somewhere. Anyway.
I was just suggesting that you could eventually share with us criteria you trust in for an EA to be valid in live sessions : some minds balance, some minds profit factor, some minds trade frequency, what do you consider ?
I'll be very please to read about others experience.
My amateur instincts and the inexperience in me think that , if an EA is about to learn something ,it should not learn it too well,and if its about to give some space to vagueness it should not handover the entire room.Think of the terms Optimization and Evolution as Averaging and Deviation.Also , i bet you have typed the term LBV in your inputs section without seeing it before , like most coders :)I apologize , 3rd coffee 0th sleep
(what's L.B.V ?)
Look Back Value , somehow plenty of mql programmers name it LBV or _LBV . MetaTrader minds think alike...
Oh ! Thanks for the l33t name : I did work on such a concept but I named it differently : a priori & a posteriori :p
I don't understand why back-test are the holy grail of making your choice. I use a very simple system. I offer a demo that can only trade .01 lots. If the customer enjoys the product and sees the potential in the SYSTEM then they will be interested in buying the upgraded package that lets them leverage their account to the fullest. The ease of making an EA that can just 0 pip spread scalp on a back-test and the report its exponential gains isn't benefiting anyone but the seller, and dishonestly. Offer a demo, let the buyer decide if the fish wiggles the way they like and the let them scale in. It also really helps with compatibility issues!
Backtesting are based from previous data but in real trading conditions you are dealing the future that is not yet known, an EA with good results on a 5 year backtest data wont guarantee that it will work the same in the future, besides there are other things to consider in real market, such as slippage,re-quotes, news time, etc. In short excellent results in a backtest environment will not ensure success in real live trading conditions.