backtesting EA's

New comment
 

Hello

I'm new to MT5 and never used automated trading before. I'm quite intriged by the potential and decided to makae the leap.

I've imported quite a few EA's but when I watch them perform in the Strategy Tester Window they never buy or sell at the right times. Has anyone encountered something similar. Is there something stupid I'm not doing or are the EA's just not written correctly? It would be nice to get a simple moving average cross over working Andrew

 
andyduvall:

Hello

I'm new to MT5 and never used automated trading before. I'm quite intriged by the potential and decided to makae the leap.

I've imported quite a few EA's but when I watch them perform in the Strategy Tester Window they never buy or sell at the right times. Has anyone encountered something similar. Is there something stupid I'm not doing or are the EA's just not written correctly? It would be nice to get a simple moving average cross over working Andrew

Welcome to the Amazing World of EA ;-)
 
ah I see.
cheers :-)
 

Hi,

Ah yes, don't trust the Strategy Tester. It sucks.
Build your own tester, your very own unique tester for your EA.

Test it on Excell is quite convenient.

Regards,

JHG

 
JHG:

Hi,

Ah yes, don't trust the Strategy Tester. It sucks.
Build your own tester, your very own unique tester for your EA.

Test it on Excell is quite convenient.

Regards,

JHG

Hello,

It may be interesting if you explain why/how Strategy Tester "sucks" ? or give a reference ?

Thanks

 
angevoyageur:

Hello,

It may be interesting if you explain why/how Strategy Tester "sucks" ? or give a reference ?

Thanks


Just from my experience that Strategy Tester cannot produce the same result as the actual as how it should be.

I keep on journal of the actual result my EA produced, open at, TP,SL, quantity, win/lose. And then compared it with Strategy Tester and it produce different result.
Sometimes the order is not opened, sometimes with different side, different TP / SL.
Don't know whats wrong with it, maybe my EA is too complicated for Strategy Tester to test it, since it requires to analyze every new "tick" vs previous data.

 
JHG:


Just from my experience that Strategy Tester cannot produce the same result as the actual as how it should be.

I keep on journal of the actual result my EA produced, open at, TP,SL, quantity, win/lose. And then compared it with Strategy Tester and it produce different result.
Sometimes the order is not opened, sometimes with different side, different TP / SL.
Don't know whats wrong with it, maybe my EA is too complicated for Strategy Tester to test it, since it requires to analyze every new "tick" vs previous data.

Right, if you use tick data Strategy Tester can't be very accurate. There is no "true tick" with Strategy Tester, only generated tick from OHLC prices.

This article will probably be of interest.

 
angevoyageur:

Right, if you use tick data Strategy Tester can't be very accurate. There is no "true tick" with Strategy Tester, only generated tick from OHLC prices.

This article will probably be of interest.

Oh, thanks for showing the article.

And if i'm not mistaken, Strategy Tester can "read" the next "tick" data which should be "hidden", if the EA is logically incorrect.
in example the data being analyzed currently is 10 AM, and the EA mistakenly reading the "next tick" at 11 AM (example by reading array Close[0]), reading the OHLC, calculating and giving great result of 100% profitable !

 
JHG:

Oh, thanks for showing the article.

And if i'm not mistaken, Strategy Tester can "read" the next "tick" data which should be "hidden", if the EA is logically incorrect.
in example the data being analyzed currently is 10 AM, and the EA mistakenly reading the "next tick" at 11 AM (example by reading array Close[0]), reading the OHLC, calculating and giving great result of 100% profitable !

I think you are talking of a "Testing Grail" by using 1 minute OHLC, see here for more detailed explanation (search title "1 minute OHLC").
 
angevoyageur:
I think you are talking of a "Testing Grail" by using 1 minute OHLC, see here for more detailed explanation (search title "1 minute OHLC").
Ahh yes there are many EAs in the market in here that show tests using that then expect you to pay for them with no real money account data. Those should be banned IMO. They shouldnt be allowed to sell without 3 months of real money forward account statements. Just as many crap scams as MT4.  Anyone that pays $1 for any EA without forward account statements is a complete fool. 
 
metalriff:
Ahh yes there are many EAs in the market in here that show tests using that then expect you to pay for them with no real money account data. Those should be banned IMO. They shouldnt be allowed to sell without 3 months of real money forward account statements. Just as many crap scams as MT4.  Anyone that pays $1 for any EA without forward account statements is a complete fool. 
It's difficult to take general measures as "those should be banned". My suggestion is to publish a review on product you find as "scam". All products on Market can be backtested.
12
New comment
Reason: