Why EAs and manual systems differ?

I have in the past been looking for eas that work well and ive really only found one (with 86% drawdown , hey but it works for 10 years on eurusd!) . Looking for eas that work and not finding them is very discouraging because

(1) Alot of systems out there are rather mechanical and can be programmed. If there aren't any succesfull ones, how would manually trading the system give different results (you always hear follow the rules right?)?
(2) What makes nonmechanical systems any more profitable? Advanced human visual recognition abilities?

However, I'm suspicious of point (1). Specifically, i've tried some eas for the Cowabunga system. They all fail. Yet, on the same time period the author who has used this system and has posted all his trades for 2 years, has a 59% success rate with atleast a 2:1 RR ratio! That sounds like, even with a mechanical system like the cowabunga system, programming an ea to match doing it by hand is hard. I've been meaning to look into why the cowabunga eas don't make the same trades as the trades posted by the author. The only sort of non mechanical thing about the system is the placing of the stops below the previous low, but that can still be programmed.

Anyways, does anyone have any ideas why a mechanical system would differ from a trader trading the same system. Is it because the trader would add more rules, or somehow not use it in bad market conditions. I.e. things that aren't included in the trading system and the rules. the mysterious human element?

Because developing any EA that has a real edge would be very difficult....what you see on open forums and the like are basic strategies that are too simple to have a real edge over the long term as the market is too efficient for this. Manual strategies work because nothing can replace the human eye on what constitutes a good setup for your parameters. For example, a pinbar may look good relative to one pair's price action but not seem relevant in another scenario. How do you program that with any reliability?

Another example is drawing price pivot levels or supply and demand - these areas are difficult to programatically identify. I've tried, and had some mild success, but it always depended on parameters for the program and that varied from pair to pair. Too subjective to waste time trying to program for.

After 5 years of trading manually, which is very boring, I started to look into automation.
First thing was to automate my own manual strategy that I considered to be the most simple strategy in the world...
but what a surprise... in short, it was impossible to do. Because it was impossible to make the correct algorithms.. in spite it
was my own strategy that I had been successfully for the past 5 years.

So, now after almost 9 yeas of trading and almost 4 of them working in automation I must say that
Automation for forex has to be taking in a whole different context then manual trading.
The proof is that every single thread we read year after year attempting to code someone manual strategy it fails.
Mostly due that all manual strategies require some points of reference (indicators) but is the discretion used
that makes them profitable .... so, Any attempt to automate an strategy that uses indicators will fail 100% of the time.

The only viable possibility to make decent codes for forex are only mathematical. Nothing more
will work because everything else require discretion... and discretion can not be coded.

J.

Profitable discretion also seems to be hard to be taught! Some1 will figure it out

So in other words don't believe the logic 'forex can't be consistently profitable otherwise someone wouldve programmed it and every1 would be rich.' very hard to program a consistent winner, humans are quite good at pattern recognition/discretionary things, even with simple strategies.

I'm sure someone has programmed an EA that makes them consistent money, but it isn't anything nearly as simple as the garbage you find on open forums. These are "black box" systems and are very sophisticated, with logic that attempts to simulate some of the discretion humans can provide. That said, there's always a subjective aspect to trading, even in BB systems. Trading is more of an art than it is a rigid science, which is hard for some, including myself, to grasp as we tend to want to nail down the rules.