Midas Indicator coded by NT

[NinjaTrader_Bertrand]

You will then need to install the file manually and compile to see where the duplicates were found - it could also be it needs code changes to make it ready for NT7, this would be shown after a manual install / compile as well.

[RDPoS]

What is the procedure for doing that please?

[NinjaTrader_Bertrand]

You would need to copy the cs file contained in the zip to the Documents > NinjaTrader7 > bin > custom > indicator folder and then compile your files by opening any script in NT's programming editor and pressing F5 then. It will list the offending code lines / section then at the bottom of the editor.

[RDPoS]

Here are the active ingredients that are causing the errors with I-MIDAS...

 
sessionend = new DateTime(Bars.SessionBegin.Year, Bars.SessionBegin.Month, Bars.SessionBegin.Day, Bars.SessionEnd.Hour, Bars.SessionEnd.Minute, Bars.SessionEnd.Second);
 
if (startTime.TotalMinutes == 0 || ToTime(startTimeDate) < ToTime(Bars.SessionBegin))
startTimeDate = new DateTime(Bars.SessionBegin.Year, Bars.SessionBegin.Month, Bars.SessionBegin.Day, startTime.Hours, startTime.Minutes, startTime.Seconds);
endTimeDate = new DateTime(Bars.SessionBegin.Year, Bars.SessionBegin.Month, Bars.SessionBegin.Day, endTime.Hours, endTime.Minutes, endTime.Seconds);
 
if (startTime.TotalMinutes == 0 || ToTime(startTimeDate) < ToTime(Bars.SessionBegin))

Session Begin is the problem. Is that a change from NT6.5 to NT7?

[NinjaTrader_Bertrand]

There were code breaking changes going to NT7 in those methods, please give the attached version a try, just copy over your present cs file in the indicator folder again and then recompile everything. I can't give credit to the source since I don't remember where it's taken from but seems to work fine for me.

[RDPoS]

It looks like all of our MIDAS problems have been solved by the Gentlemen at Ninjatrader.....

http://www.ninjatrader.com/support/f...ighlight=midas

[koganam]

I found a flaw in the MIDAS Explorer "MIDAS Touch" indicator when launching from a range chart.

Using a 3-range chart on ES from today, I launched 2 MIDAS curves. A red curve from the peak at 18:59:33, and a green curve from the minimum corresponding to the 8th bar of the timeframe 20:12:12.

Note that MIDAS launched the curve from the 1st bar of the 20:12:12 timeframe, and not the 8th, yielding an incorrect curve result.

The fix via NT would be to further subdivide bars to actually have a count, when there are more bars generated than 1 per second. So the first bar of 20:12:12 would be called 20:12:12:1 and the 8th bar of 20:12:12 would be called 20:12:12:8.

Not even trying to fix it to the low will work as it just sees the first bar of any one second interval.

Any way to correct this limitation?

In the meantime, why not fix it to the bar after the low bar. That would give a more reasonable anchor point than what you show.

But, it is a flaw in the code. Why not ping the developers? They have to anchor off the last bar if bar times are the same, not the first bar; that is only reasonable. After all, the last bar in the chain IS the most recent bar with that time stamp. It should be a simple fix.

Maybe we should open a separate thread for this, as it is an important issue, and really a different indicator from the one that this thread was discussing?