LinRegSlope in Degrees

November 21st, 2010, 01:19 AM

I'm looking for a way to have the Ninja Indicator LinRegSlope expressed in degrees so I can run some tests across a basket of instruments. Before I get too carried away has anyone already done this or seen this?

Default LinRegSlope code is:

Code

 
protectedoverridevoid OnBarUpdate()
{
double sumX = (double) Period * (Period - 1) * 0.5;
double divisor = sumX * sumX - (double) Period * Period * (Period - 1) * (2 * Period - 1) / 6;
double sumXY = 0;
for (int count = 0; count < Period && CurrentBar - count >= 0; count++)
sumXY += count * Input[count];

Value.Set(((double) Period * sumXY - sumX * SUM(Inputs[0], Period)[0]) / divisor);
}

And the output is related to the tick size therefore you can not directly compare ES and 6E, for example.

November 21st, 2010, 01:30 AM

see this NinjaTrader Support Forum - View Single Post - Slope function output

November 21st, 2010, 01:36 AM

nexusfi.com (formerly BMT) CollectiveMA also has slope code built in, easy to modify for your purpose.

BMT Collective MA

Mike, just curious what exactly the futures.io (formerly BMT) Collective MA is? I apologize if this was described elsewhere. I saw it was part of the new Envelope indi but I had never seen it before then. Thanks.

Blz

Mike

November 21st, 2010, 02:12 AM

Thanks! And talk about a fast response!

Testing to see if this is correct:

Code

Value.Set(System.Math.Atan((((double) Period * sumXY - sumX * SUM(Inputs[0], Period)[0]) / divisor)/TickSize)*180/Math.PI);

November 21st, 2010, 02:50 AM

Actually if anyone can check that code it would be great. My weekend hypothesis testing is the "Euro trends more than the ES" thing that came up in another thread. One way I thought to test that was to look at the distributions of LinRegSlope for the two instruments, but that will only work if it is true apples-to-apples math.

Preliminary histograms attached. 20 days of 1 min data.

LinRegSlope in Degrees

Discussion in NinjaTrader

LinRegSlope in Degrees