How many years should I test My system, to consider it, profitable?

I have posted yesterday about my system that if traded entire 2007, would double my account, but I have only backtested the system in 2007 and 2006, every year account doubled, my question if for how many years should I backtest it, to be sure that I have a winning system, I have no experience in how many years I should test it, are 2 enough? should I test 4 or 6 years? I don't have a clue about that I hope someone can help me here.

Thanks Mike

Hello mike, interesting question.
I will say this...that you must realize that now one will give you the right answer on this question. Someone are testing them for 2 years, some for 3....but all people are scared of system that it works for 1year and when they start to use it for real money, it dont work anymore!! This happen to a lot of traders with the trading robot...So sometimes its better to start using it as sson is possible to start making the profit before it want work anymore.

So it`s very hard to give you the right answer, anyway personaly i would test it for 2 years minimum....But in this time you can learn a lot of more principles of trading...

i hope that it helps a little

The standard should not be based upon how many years, it's how many trades.

I've heard the number 100 trades being floated about. Some even go as far as 1000.

100 years before you risk a dime ...

ForexMoments is correct, it's the number of trades you consider 'statically significant', this could be anywhere from 40 upward depending on which techniques you use to evaluate the results. You need to make the same decision on sample size to determine when a system stops working also.

Stats or no stats, tis nothing more than a curve-fit if price is the only input.

Hi Mike,

Forexmoments is correct it's more about the number of trades in the backtest rather than an X amount of years...If you are trading from a 1 min chart I would say 2 years is enough..

What system isn't a curve fit to some extent?
I think the term you need is 'over-fitted', these are of course, useless.

Hi,
When i started on this long road called Fx trading i
remember reading a posting that said if you could make
a profit for 6 CONSECUTIVE months, staking the
absolute minimum, then you could consider yourself
to be on the right path to making consistent returns.
My best at the moment is 3 months - +12%, +12%,
+8%, then crunch time with trades hitting my SL.
So i've still got to aquire more knowledge and
experience- it's a long road.
D.
Ps - paper trading your method/system is ok- but until you start trading hard cash, even
a small stake,you won't know how consistent you'll be, until you overcome the
emotional maelstorm which kicks in when your money is going down the pan!!!.

Hi, and thanks everyone for the replies, Its a medium term strategy have made 400 trades in 2 years.

I would first backtest it a few years, and before I start trading some real money I would run it for a while on a demo, or a live account with a little bit spare money..
I wouldn't rush into it to soon, check out all the negative things and try to fix them.

Your two years of back testing already showing 100% return. Now even if you make 30% as a worst case scenario , it is still good, right?BTW, could you share the system that shows 100% for the past two yearsThanks

i would backtest as much as far back as you can. I would try it on a demo account for at least six months too. Most successful backtested strategies fail when they are forward tested, because like tdion said, they are just curve fitting.

I would test it forward since you have confidence in it. Just don't get crazy with position sizing. Then while testing forward you can get the emotional side of this game working for or against you. Or like mentioned run demo account and test it going forward.

3 years minimum. 5 or more is ideal

I would at least test it with more recent years... ranges, volatility etc have changed over the past 10 years. If your system works as well in 2016-17 as it did in 2006-7, then go demo for a couple of month and all should be fine.

i would test the system or the market aproach on 300 to 1000 trades on the simulator. demo trading the same, the higher number of independent trading decisions the better.

Ten years documented demo-trade history should give you an idea as to whether you might succeed.

Anybody that says a set amount of years doesn't know what they're talking about.
You don't need an X amount of years.

What you need to do is test your strategy over a period that has different market conditions as well as a large enough sample size.
Usually 4 years does that for me on the 1H timeframe. If you're testing a higher timeframe you might want to go longer and if you're doing a lower timeframe you might get away with less but the amount of years should not be your focus. More is always better of course.

So to summarize:
-Different market conditions
-Large enough sample size
-The longer period you test the better your insight will be.

It's as easy as that.
Don't listen to anybody that just says "10 years bla bla bla"