# How much period we should test before we start using that EA in real Account.

1484

How much period we should test before we start using that EA in real Account.
Documentation on MQL5: Standard Constants, Enumerations and Structures / Environment State / Account Properties
• www.mql5.com
Standard Constants, Enumerations and Structures / Environment State / Account Properties - Documentation on MQL5
Moderator
18056

kchandek:
How much period we should test before we start using that EA in real Account.
3 or more years in 4 or more symbols without optimization . . .  to demonstrate consistency.  All in my opinion . . .  DYOR.
3378

RaptorUK:
3 or more years in 4 or more symbols without optimization . . .  to demonstrate consistency.  All in my opinion . . .  DYOR.
5390

kchandek: How much period we should test before we start using that EA in real Account.
Perform a Back-Test upon 5+_Years and 5+_Symbols...........Perform a Live-Test for at least 3-Months. Strategy_Optimizer leads to Curve-Fitting. All in my Humble Opinion.
3378

Ubzen:
Perform a Back-Test upon 5+_Years and 5+_Symbols...........Perform a Live-Test for at least 3-Months. Strategy_Optimizer leads to Curve-Fitting. All in my Humble Opinion.

5 yesrs back testing... what is the difference between 5 years and one year? why do you need such along period? and then live testing for at least 3 months, why not six months?

Curve fitting is a mathematical term? Thank you.

5390

PCWalker:

5 yesrs back testing... what is the difference between 5 years and one year? why do you need such along period? and then live testing for at least 3 months, why not six months?

Curve fitting is a mathematical term? Thank you.

5 minus 1 is a difference of 4 years. At least 3_months so 6_months will qualify also. I'll be waiting for your mathematical method to the poster's question .... thank you.

3378

Ubzen:

5 minus 1 is a difference of 4 years. At least 3_months so 6_months will qualify also. I'll be waiting for your mathematical method to the poster's question .... thank you.

I was asking a question about the term, I have just started studying a little bit of statistical optimization in R (Self study) so I can follow the ideas and good advise to succeed in analyzing expert advisors.

Thank you.

Moderator
18056

PCWalker:
It is based on practical experience of running many EAs in the Strategy Tester on many different Symbols.  Optimization is mostly curve fitting.
5789

RaptorUK :
3 or more years in 4 or more symbols without optimization . . .  to demonstrate consistency.  All in my opinion . . .  DYOR.

I agree with raptor, but maybe a lot depends on what trend is the expert going to follow. For the primary trend three/four years may be sufficient. If you are following  secondary trend or an intraday trading or scalping less than 3 years can be also fine.

Also very important is Time Frame. there is for example a big difference between a 15minute time frame and an 8H one. I mean that minimum testing period would be much different.

5390

PCWalker:

I was asking a question about the term, I have just started studying a little bit of statistical optimization in R (Self study) so I can follow the ideas and good advise to succeed in analyzing expert advisors.

Thank you.

How Not to Fall into Optimization Traps? - MQL4 Articles
• www.mql5.com
How Not to Fall into Optimization Traps? - MQL4 Articles: strategy tester
Moderator
46443

kchandek:
How much period we should test before we start using that EA in real Account.

For me it's not just a question of paper trading time, as you can use real account as live testing.

My method is wait the enough time to find a good performance synchronism between backtesting/forward testing/live paper trading testing and small lot size live real trading testing before start to increase real account volume.