Dynamic EAs - page 3

 

Grail

My friend upon intial backtesting I have not seen any EA as good as this one, I have tested nearly over one hundred, all pretty much rubbish, early days yet but this could be the holy grail I have been looking for, keep up the good work.

 

Modification

Forgot to ask are you able to install Magic Number for both the EA's and if you are capable a Stealth Mode setting or Hide Take Profit and stop/loss, thanks again will keep you posted on results, will finish my testing this weekend if all good will go live on Monday.

 

Ea mod

Silly me I see you have magic number already, hopefully you can install stealth mode?

 

Thanks !

Extremely nice work ! Thanks, MrTools. Green pips for you !

 
john1368:
Forgot to ask are you able to install Magic Number for both the EA's and if you are capable a Stealth Mode setting or Hide Take Profit and stop/loss, thanks again will keep you posted on results, will finish my testing this weekend if all good will go live on Monday.

John and Mikhael thank you all for your kind words!

Will work on stealth mode and John by going live i hope you mean demo first or maybe small real account, till you are sure about it, because back testing can be very deceiving most of the time!

 

Backtesting

Having so much trouble with backtesting, always getting varied results, wish Metatrader or someone else made a stable reliable Expert advisor backtesting software that gives accurate results, to many variables, only way is to forward test I think unless someone has a better idea?

 

DoubleMacd+DssBressert+nlr+timefilter-1a EA.

H1 timeframe, IBFX broker.

===============

All the versions of the Dynamic EAs - see this post.

Original thread about all Dynamic EAs is here.

Dynamic DoubleMacd+DssBressert+nlr+timefilter-1a EA with all the files to be used: here.

Files:
 
john1368:
Having so much trouble with backtesting, always getting varied results, wish Metatrader or someone else made a stable reliable Expert advisor backtesting software that gives accurate results, to many variables, only way is to forward test I think unless someone has a better idea?

Different results in backtesting is because different bars in test (Metatrader is downloading the history during the backtesting). Besides, modelling quality = 90% is reliable only.

I am forward testing as you see from the statements.

 

About Back Testing

To ND

I use forward testing only.

Is 90% modeling reliable enough ?

I thought Back Testing using only one time frame.

and back testing when the market is closed, Spreads are very different.

If EA is MTF, is it still reliable ?

for example, Let's say EA decide trade direction with 4H time frame.

Then back test doesn't trade properly. I thought.

Or Did I miss understand the back testing.

 

If EA is MTF so backtesting is not reliable. It is not reliable if EA is using open bar coded (such as 0 bar), or any high/low of close bar (Firebird EA for example), or any combination of it.

That is why - when the people are selling EAs and making the comclusion based on backtesting resuts - so hey are no telling the true: because, if we can not see about what is inside the code as mql4 so any backtesting reslts can not be trusted.

Besides, many EAs are performing differently in forward testing just because of the above mentioned reasons.

Other reason - some indiators using inside the EAs.

Example: Brainwashing EA using iTrend: the value of iTrend indicator is different for different brokers (on closed bar).

The other example - old versions of public BrainTrading indicators: because of spread using in calculation.

Same problems are for any MAs crossing EAs of MA is using as 'Apply to' Open, High, Low, Median, or LWMA and so on. Example: MAChannel EA.

So it is the general problem/issue of Metatrader's backtester.

Reason: