Hello. I have been struggling with this matter for two days now, and can't seem to find of any solution about it.
No matter what I do, the values that I get from any custom indicator during backtesting of EA are always totally different from what appears on the charts. I have updated the M1 historical data from Alpari, etc.. Here is the function that I am using:

init(){
if (LocalTime()>StrToTime("7:30")&&LocalTime()<StrToT ime("7:45")){
Print(DoubleToStr(iCustom(NULL,PERIOD_M5,"Juice",0 ,0),8));
}
}

Running the EA on M5 timeframe.

Here is a piece of the results:

2006.06.03 19:21:21 2006.06.02 07:34 ICustomTest EURUSD,M5: 0.00000000
2006.06.03 19:21:21 2006.06.02 07:34 ICustomTest EURUSD,M5: 0.00000000
2006.06.03 19:21:21 2006.06.02 07:33 ICustomTest EURUSD,M5: 0.00000506
2006.06.03 19:21:21 2006.06.02 07:33 ICustomTest EURUSD,M5: 0.00000506
2006.06.03 19:21:21 2006.06.02 07:33 ICustomTest EURUSD,M5: 0.00002318
2006.06.03 19:21:21 2006.06.02 07:33 ICustomTest EURUSD,M5: 0.00002318
2006.06.03 19:21:21 2006.06.02 07:33 ICustomTest EURUSD,M5: 0.00000000
2006.06.03 19:21:21 2006.06.02 07:33 ICustomTest EURUSD,M5: 0.00000000

I can clearly see that on the charts Juice shows values as high as 0.0003. Results seem to vary no matter how big values are. I get the same strange results with all custom indicators I've tried so far...
Has anybody encountered such a problem before? I'd be really glad if you let me know what to do.
Thank you in advance!

A few minutes ago I replaced the iCustom Juice call with a iStdDev (=Juice?) one, and got the appropriate values. So it seems that I am not addressing Juice right, but what is my mistake?

Juice have 2 parameters. Look at the code


extern int Periyod=7;
extern double Level=0.0004;

and two buffers
#property indicator_buffers 2
So u have to pass the parameters to the iCustome command
double J1 = iCustome(NULL,0, "Juice", 7, 0.0004, 0, 0);
the above command will return the first buffer value from the current bar

Hope it will help
Eli

Thank you for replying! I just found out what I should do - read the value with shift=1, as during the backtesting the shift=0 value sometimes varies too quickly with a very large amplitude. Again, thanks a lot!