When it comes to automated trading software you have to be careful and do your research. I wrote an article a few days ago regarding MetaQuotes had how they recently opened up a marketplace specifically for MQL4 expert advisors, which is great.
But with great power comes great responsibility and if you're not careful you could find yourself losing more money than you're prepared for.
One of the things that you absolutely must do when you're looking at a potential EA for your portfolio, is to test it. I don't mean just run a test in strategy tester and then jump straight to a live account.
Doing that would most definitely be financial suicide!
You need to run a few tests in Strategy Tester and play around with various settings and optimizations. See which settings seem to yield the best results. You're not just looking at profitability but also keep a close eye on your draw down.
From there you'll need to find a decent VPS provider and install the expert advisor on one or more charts. Ideally you want to have a few versions running on a number of different pairs.
Also try to have a range of settings that you can be testing over the period.
The advantages of using a VPS are many but the most obvious one is the 24-7 connection to the Internet. If you're home connection is anything like mine then it's definitely not wise to risk running an EA on your personal computer.
There are just too many things that could go wrong and potentially skew your results. A losing expert can look profitable and a winning one can look like a loser. If you're serious about running automated trading software then you should get a quality VPS.
They're quite cheap and I have a couple that I use. I currently have a Windows VPS with a guy and the cost is peanuts really. Only $18 a month and the hardware spec and reliability is pretty impressive.
Definitely beats my laptop, which isn't too shabby at all, but my Internet connection just plain sucks! One of the many 'benefits' of living in 'paradise' 🙂
What's An Ideal Period For Testing
Many of my students ask me how long they should test an EA before going live. My answer to them is always the same and I'm afraid it's not very sexy at all...
... as long as possible!
It's important that you feel as confident as you possibly can with your automated trading software because after all, you'll be risking your hard-earned money with it.
For me I'd look to test for at least a month, maybe even two.
After that you can continue to test on a live account but again, a word of caution.
Don't rush into it and gamble everything you have right from the start.
Be smart about it. Start of with a micro-account and have the smallest lots that your particular MQL4 expert advisor allows. Again, you'll want to leave this running on your VPS for a decent period of time. I'd recommend a month at a minimum but again, use your own best judgement and be smart.
During your testing phases you should also be monitoring the experts functionality. Now again this may seem boring and very un-sexy, and it is... but you simply have to do it.
Watch out for any errors that might be printed to the 'Experts' tab of your MetaTrader4 terminal. You can usually ignore warnings but anything marked with a red triangle deserves your attention and you should take note.
This point is something I learned the hard way.
When I first started coding in MQL4 back in 2007, I'd written what I thought was a pretty good EA. I'd tested it thoroughly in Strategy Tester and it didn't present any errors. I was super excited to let it lose on a live account and in hindsight, maybe too excited.
Well the very first night I left it running and went to bed, I woke up to find my whole testing account of $3,000 completely gone. It was a testing account but it was live, with real money. I was sick to my stomach and I couldn't believe what had happened.
I thought I'd tested everything, and I had, but it was one of those 'black swan' events. To this day I'm not quite sure what it was but my best guess is some kind of server or broker error. Maybe a disconnection or an interruption in the data.
Whatever it was, the fact is my EA didn't check for it and had no way of handling it. So it just threw up it's hands and threw away my cash.
Oh well! Lesson learned and I no longer make mistakes like that. I've learned from that first newbie error and hopefully my loss will serve as a warning to you.
When dealing with automated trading software there's three things to remember...
...test! test! test!
Head over to the MQL4 marketplace and grab a few expert advisors and start testing them. If you're feeling particularly adventurous you can even try coding your own.
Best of luck and take care out there.