The .zip file is here: https://www.mql5.com/en/forum/173010
What indicators do we use from the zip folder you posted? There seems to be a number of indicators and templates.
Also,. what indicator is the Moving Average indicator?
Indicators used : 3c_JurX and 3c_JRSX, settings as default.
Timeframe : 4 Hrs
Entry rule : BUY when the MA on chart changes color to Yellow and also JRSX histogram must confirm as a Violet bar. ONLY AT CLOSE OF A BAR.
SELL when the MA changes color to RED and also the JRSX histogram must confirm as Majenta bar. ONLY AT CLOSE OF BAR.
Exit rule : When EITHERof the MA and JRSX histogram changes color. ONLY AT THE CLOSE OF A BAR.
Visual backtesting looked good so far. Actual backtesting could not be done in the absence of an EA.
The good part of the system is not over-optimised at all hence should work in any market conditions.
I created a template for the tester using these indicators and watched things develop in realtime, the oscillator indicator does not seem to work at all well when running in the backtester, I would suggest that needs sorting out before anybody tries to write an EA.
Can you send me those indicators?
Files are here https://www.mql5.com/en/articles/1450
The indicators are 3c_JurX and 3c_JRSX, check my first post.
The moving average is 3c_JurX.
What can be wrong with the RSX? Why it doesn't work in backtesting as per you? Your input is highly appreciated.
The idea for using the histogram is like this. A bar is violet only when its value is higher than the just previous bar and reverse for majenta. Similarly the MA is yellow when its value at the close of 4Hr Bar is higher than the previous price bar and opposite for Red color.
So for opening a position the EA has to identify the MA color and the current histogram value and previous histogram value.
I'm not too sure, the code is very complex, I assume the backtester is exposing some kind of numerical instabilty. This is a shame as this set of indicators uses a lot of common code, so one must assume they all have a high chance of doing the same thing. If you assume the indicator is correct, then what does that say about the backtester?
Thanks for the link to the files, but where are they? I do see alot of code, but is there a plce to download it all?