builtin iSAR Indicator Function returns different/wrong Values when back tested as a Script, but not as an EA

[[Deleted]]

hi all,

as I'm currently working with an iSAR-Strategy and willing to back test it in the strategycenter, I run in some inconvinience:

Comparing the values returned from

iSAR(

string symbol, int timeframe, double step, double maximum, int shift)

once called within an indicator, once called within an EA - they missmatch (I use an export to CSV files for comparing).

If the function is called from an indicator, only the values of the selected period are correct, while other periods return static values.

If the function is called from an EA, everything works fine.

I looked around in the documentation and the book, but couldn't find anything.

Question 1: Is it possible code executed as indicator on back testing doesnt has access to correct values for different periods then the back tested period?

Possible Question 2: if Q1 is true, does anyone know a workaround for this? (Despite it wouldnt make to much sense to me, as back testing with EA returns correct values - with indicator not)

Thank you for your helping answer.

p.s.

you can verrify this behaviour with the files attached, if you like to.

[William Roeder]

zeitgeist23:

Question 1: Is it possible code executed as indicator on back testing doesnt has access to correct values for different periods then the back tested period?

Tester limitation, can't access Bar 0 from other TF/pairs.