Sti - page 2

New comment
 

sti

It doesn't seem to go onto the graph. Does it need to be modified? Nothing comes up.

 
tomsd:
I have created an EA that seems to work great, but it has a problem when the news come out.

Maybe somebody can help to create some filter or something like that.

Thanks

First: Thank you for sharing an expert with the community. It is actually quite sophistically programmed.. Respect.

Now the downside. Backtesting is already quite weak, when you have 90% modelquality, but utterly unreliable with <50% like you have. The reason: You have large gaps in the data. In this gaps can happen anything: SLs get hit, TPs, Signals are made, but you dont see them because for MT4 they are not existant.

I just tried you expert in the 4 majors in the timeframes M15,M30 and H1. I have >80% quality for these for the time since 01.01.2004.

Result: The expert just burns money. That is not just a question of a filter, since a filter only makes sense, when you have good and bad streaks, but you do not have streaks. 3 of 5 trades win completly random, but they win only with about 10% of what the loosers loose = ruin. You can reduce the SL to reduce that, but then you suddenly have 4 out of 5 trades loosing (though SL), and that before even get close to a loss win ration of 50%/50%.

So ... back to the drawing board. If it is a comfort, I have made the same mistake like you more then once, until I understood the issues of the dataquality.

 
kasmodiah:
So ... back to the drawing board. If it is a comfort, I have made the same mistake like you more then once, until I understood the issues of the dataquality.

That's the holly truth of backtesting - U should have a good dataquality and if U test EA in metatrader do it on m1 frame based on every tick - that gives U the best modeling probability.

 

Hi

Does anybody have tested STI? It's making profitable trades but something still missing.Please make some suggestions how to make it better.

Thanks

 

Here is demo account statement for STI. On EUR/USD M15 it looks great, isn't it...

Files:
detailedstatement_28.htm  135 kb
 

Hello,

Not bad of the whole, I say to follow with interest

 

I have some problems to make backtest. When I made it first time then modeling quality was more than 90%, then I changed inputs and modeling quality became 34%. Now I can't get modeling quality back to 90%.

Does anybody know what to do?

 

It would be nice with some arrows on the chart when you backtest.

Can someone fix that?

Thanks

Smeden

 

Tomsd,

I think you have use TP 6 and SL 50.

If we go for TP 10 and SL 50 Result will be more better

It is only my suggestion.

 

One More back testing i did,

USD/CAD TP 10 SL 60, 1 Hr chart.

results are very good

123456789...22
New comment