Estimated time remaining 900 hours. What!

 
Hi we've created this EA coming soon at fxautomated.com that we think will be the best ever in the market. We are currently optimizing it on a powerful server that is on 24/7/365 on but the estimated time to complete is 900 hours thats about 35 days. Can you guys please create a cloud network for mt4 to speed up strategy testing like in mt5.
 
Maybe you should optimize your code first.
 
its day five now and almost 800 hours remaining so the time seems accurate since we use quality historical data
 
tonny:
its day five now and almost 800 hours remaining so the time seems accurate since we use quality historical data

Please define quality historical data, how long is it, is this a tick data, and what is the size ?

I've collected EURUSD tick size before, couldn't stand the size, had to aborted it.

Maybe you just have to consider, that this, is a 1 month forward testing ;)

 
RaptorUK:
Maybe you should optimize your code first.

Did you forget to enable Genetic algorithm? If it shows more then 11000 passes, you did.

Optimize your code first.

If you're not waiting for price to reach a specific level (to open or to move stops, etc,) wait for a new bar.

If you're waiting, don't recalculate everything each tick.

From my code

double  ca.below,   ca.above=0;                 // Import from CallAgain
int     start(){                                                
        #define INF 0x6FFFFFFF // Not quite infinite, Jul 2029, or 1,879,048,191
    if (Time0 != Time[0] || ca.below > Bid ||   ca.above < Bid){    // Important
        Time0 = Time[0];    ca.below = 0;       ca.above = INF;
        :
}
void    CallAgain(double when, string why){
        if (ca.above > when) if(when > Bid - Point_2)   ca.above = when + Point_2;
        if (ca.below < when) if(when < Bid + Point_2)   ca.below = when - Point_2;
}
:
    if ((Bid - openBid)*DIR < 0){ CallAgain(openBid,"opn"); return; } // Wait for price
    OrderSend(...)
 
onewithzachy:

Please define quality historical data, how long is it, is this a tick data, and what is the size ?

I've collected EURUSD tick size before, couldn't stand the size, had to aborted it.

Maybe you just have to consider, that this, is a 1 month forward testing ;)


I changed maximum bars in history setting and made a huge download of data(like over 100MB to estimate) now strategy testing is more strenous to the computer but as i understand this is the most accurate way to optimise as it returns more realistic historic performance. If there were a cloud network for mt4 it would have been great because ive tested the one on mt5 and it works great, a days testing is completed in about 10 minutes but im not sure of the accuracy mt5 strategy tester because it doesnt have the option of downloading more history data.
 
WHRoeder:

Did you forget to enable Genetic algorithm? If it shows more then 11000 passes, you did.

Optimize your code first.

If you're not waiting for price to reach a specific level (to open or to move stops, etc,) wait for a new bar.

If you're waiting, don't recalculate everything each tick.

From my code


I want the most accurate result possible not just speed everything up then start selling a losing EA like what most sites do.
 
tonny:

I changed maximum bars in history setting and made a huge download of data(like over 100MB to estimate)
That's not huge . . . try grabbing a few years of tick data . .
 
RaptorUK:
That's not huge . . . try grabbing a few years of tick data . .

Yer on a fast connection its not a too huge download but in strategy testing it increases the work two folds and the time to complete by about 4 times i want the most accurate results im just petitioning for a mt4 cloud.

 
tonny:

Yer on a fast connection its not a too huge download but in strategy testing it increases the work two folds and the time to complete by about 4 times i want the most accurate results im just petitioning for a mt4 cloud.

I don't see an MT4 cloud happening . . . it's surprising any new dev is happening at all on MT4 . . . your best results will come from optimizing your code . . . but as you are doing this commercially and making money at it I guess you have done that already.
 
tonny:

I want the most accurate result possible not just speed everything up then start selling a losing EA like what most sites do.
What in my post would change accuracy?