I'm not sure if you know it - but there features in MT4 that are considered to be directed against a trader, but for a profit of a broker.
So many people think MT4 not very attractive platform, if you want to win...

That's a link that mentions this.
Foreign currency trading is an easy way to lose money - Los Angeles Times

@DayTrader : My perspective is, that we have to deal with 3 questions

1.) Are there profitable system ? I did code some which are profitable for a limited period , max 1,5 years . I am working on a system which shall be profitable for several years . I know people which informed me, that they have a limited collection of systems , which are "some" profitable for 10 years plus . I have seen there fully vaildated "myfxbooks" , no more no less .
There are NO HOLLY GRAILS , so I am being told by people I trust . If there were, I would not like to have them . I enjoy working and I would get crazy ideas with lots of money .

Thatīs my best reply to you w/o knowing which duration and which profitability is targetted , when you asked .

2.) Broker :
2.1. ) There are coīs which do not deserve to be rewarded with our cash . Some of them also use MT4 . Search the www for the names , easy job .
2.2. ) I trust Pepperstone and MBTrading enough to give them some of my riskcash and run MT4 with them . InteractiveBroker has the rest of my trading cash .
2.3. ) My automated system run slow , H1 plus for signals , M5 for ordermanagment (TSLīs)

3.) MT4 : With precondition 2.2.) & 2.3.) I love Metatrader . Prefer it before Amibroker & MC (own both) .
3.1. Reason : Once you learned it fully, you will have hugh support in the web and the software allows you so much control on anything, that there are no limits but always solutions for your requirements.
3.2. It does not cost you one cent . First ever exception to me, were a product at zero cost serves best (to me and precondition 2.2.) once again )
3.3. You have little worry with the order sync Metatrader <==> Orderserver , Your SL lieīs at the orderserver of the broker . If the web brakes down, the electricity fails , your children switch off your homeserver ...the SL is there and protects you from desaster .
3.4. Lots of buildt in solutions for storing data , very helpfull if you wish to run data analisis outside of Metatrader (ie profitability analisis in excel returning information back to Metatrader for improved settings and lots of other things one could imagine)

Some thoughts : My passport states that I am a veteran of age but that is not true . I have a technical background and did code 30 years ago intensively , than stopped for 25 years and now started 4 years ago again to code . It was not easy , also my free time is limited , but now I made it . It is sort of my "back up" planing if other things develop less nice .
In case you have "Zero" coding-experience you will need lots of "like for the matter coding" , some patience and time available . But coding is logic pure , no more no less . I love it .

My approach is, to automate what could be automated . There are "things" in trading , which cant be automated (I canīt ) . Solution : Go flat and alert the coder . Plain easy . So , whats the deal if you have to tweek manually each second week ? To me itīs fun . I do not need to be in paradise island whilst my little PC does the trading . It would take some of my trading-joy away from me .



Lobo

YES it is possible

BUT there are a few things to verify first :
- find good data. The one coming with Metaquotes or many brokers are not reliable (I guess to give more profitable backtests)
- verify and modify the spread before backtesting
- just a few Metatrader brokers are honest and don't change the spread, delay your order or trade against you. So find the good one
- once you have a profitable strategy test it live on microlots first. Be careful with scalping strategies that don't work often on real account
- use a VPS or colocated server

Not only is it possible to program profitable EA's but it's becoming more and more common as traders become more comfortable with mt4's limitations.

I have traded an EA profitably for the past 2 years, granted the drawdowns are larger than if I were able to trade it manually, but the point is I would rather make money than be right all the time.

I have noticed though that strategies do go through phases where they just stop working. It's the burden of the backtest to decide when and if to start it running again.

We talk often about the art of trading, well even in automating the process there are still decisions to be made with intuition and that is why some traders make money even when they leave the process up to their "robots" while others are in an endless loop of "changing horses."

peglegtrading 

Not only is it possible to program profitable EA's but it's becoming more and more common as traders become more comfortable with mt4's limitations.

I wonder if you have the numbers to back that bold claim.