Need ConnorsRSI

[lang0477]

Greetings,

Can anyone share a version of ConnonRSI indicator with me?
looks like the only place is BigMikeTrading but I don't have elite membership..

thanks
Billy

[NinjaTrader_PatrickH]

Hello Billy,

Thank you for your post.

Unfortunately, I am unaware of any other resource for this indicator. One of our forum members may have such an option for you.

[LarrikinDave]

Did anything come of this???

http://www.ninjatrader.com/support/forum/showthread.php?t=58423

I'm interested in the ConnersRSI too.

D.

[Harry]

Available here for members:

https://www.bigmiketrading.com/elite-circle/25094-connors-rsi-new-indicator-strategy-2.html#post346767

[photog53]

ConnorRSI

The code for this version of the RSI appears to be "open source"... and I have found no copyright restrictions on it. The versions I have seen are not coded for NT....but If you give me a couple of days (to do this in my spare time)....I will convert a version to NT and post it in the Ninja Downloads section.
IMPORTANT: The ConnorRSI has a default "rankPeriod" lookback value of 100....which means this thing will be a real CPU hog if you run it with "CalculateOnBarClose=false. There are probably ways to code around this...but I am not investigating them....I am simply converting the existing code.

The code should run fine on intraday charts....which seems to be what most retail traders favor....but I suggest leaving COBC=true.

[Harry]

The code for this version of the RSI appears to be "open source"... and I have found no copyright restrictions on it. The versions I have seen are not coded for NT....but If you give me a couple of days (to do this in my spare time)....I will convert a version to NT and post it in the Ninja Downloads section.
IMPORTANT: The ConnorRSI has a default "rankPeriod" lookback value of 100....which means this thing will be a real CPU hog if you run it with "CalculateOnBarClose=false. There are probably ways to code around this...but I am not investigating them....I am simply converting the existing code.

The code should run fine on intraday charts....which seems to be what most retail traders favor....but I suggest leaving COBC=true.

The version that I have coded can be used in COBC = false, unless you put hundreds of them on the market analyzer. The graph attached shows a test that I have run with the Connors RSI added to a chart with setting COBC = false. I have used market replay with speed x500 on a single chart for ES 03-15.

As you can see, the Connors RSI does take a little toll on the CPU during the news release and at the open of the regular session. But half of that CPU load is created by NinjaTrader itself. Let us say that it takes 30% of one of the 4 kernels of my CPU (Xeon comparable to i7) at speed x500, when referring to peaks only. Then it should not take more than 0.1 percent of the CPU capacity when used at normal speed.

That means that it should work, if you do not put more than 100 Connors RSIs in your active workspace at the same time.

Actually the code for finding the rank is pretty simple:

rank = 0.0;
            for (int i = 1; i <= barsAgo; i++)
                if(rateOfChange[0] > rateOfChange[i])
                    rank = rank + 1.0;
            percentRank = 100*rank/barsAgo;

which explains that the CPU does not get excited.

[photog53]

I converted the code...a version is available on the NT downloads

[Harry]

I converted the code...a version is available on the NT downloads

Actually, you have not converted the code, but just copied my version of the Connors RSI, and then renamed a few variables. I don't mind that you copy my code, but you should not pretend to have the indicator converted to NT7.


Comparison:

Code I published open source July 26th, 2013: https://www.bigmiketrading.com/elite...tml#post346767

if(CurrentBar < 1)
{
   streak.Set(0);
   ConnorsRSI.Set(50.0);
   return;
}    
barsAgo = Math.Min(rankPeriod, CurrentBar);
rank = 0.0;
for (int i = 1; i <= barsAgo; i++)
if(rateOfChange[0] > rateOfChange[i])
      rank = rank + 1.0;
percentRank = 100*rank/barsAgo;
if(Input[0] > Input[1])
      streak.Set(streak[1] < 0.5 ? 1.0 : streak[1] + 1.0);
else if (Input[0] < Input[1])
      streak.Set(streak[1] > -0.5 ? -1.0 : streak[1] - 1.0);
else
      streak.Set(0.0);
ConnorsRSI.Set((percentRank + streakRSI[0] + defaultRSI[0])/3.0);

Copied indicator which was published today:

if(CurrentBar < 1)
  {
         ids_Consecutive.Set(0);
         ConnorsRSI.Set(50.0);
         return;
  }    
  barsAgo = Math.Min(rankLookBack, CurrentBar);
  rank = 0.0;
  for (int i = 1; i <= barsAgo; i++)
  if(rateOfChange[0] > rateOfChange[i])
     rank = rank + 1.0;
  percentRank = 100*rank/barsAgo;
  if(Input[0] > Input[1])
      ids_Consecutive.Set(ids_Consecutive[1] < 0.5 ? 1.0 : ids_Consecutive[1] + 1.0);
  else if (Input[0] < Input[1])
      ids_Consecutive.Set(ids_Consecutive[1] > -0.5 ? -1.0 : ids_Consecutive[1] - 1.0);
  else
      ids_Consecutive.Set(0.0);
  ConnorsRSI.Set((percentRank + clbRSI[0] + defaultRSI[0])/3.0);

Only difference: A few variables were renamed. No value added.

I have not put the indicator into the download section of any forum, because I have never contacted Connors Research to ask for permission. On the other hand they have published the formula and it is a good publicity for Larry Connors that we discuss his indicators here!

But credit to whom credit belongs. Here is the link to Connors Investment:

http://info.tradingmarkets.com/ConnorsRSI_ConnorsRSI.html?utm_source=TMACRSIpage&utm_medium=textlink

You can sign up for a free copy of "An Introduction to the Connors RSI".

Just to be clear: I am not linked to Connors Research in any way, nor have I ever talked to them.

[ninZa]

This topic is going hot here

[Harry]

This topic is going hot here

I also took it as a compliment. In the end he did not find anything to improve.

[ninZa]

I also took it as a compliment. In the end he did not find anything to improve.

This means your code is optimal
Congrats Fat tails lol.

[photog53]

Harry....when converting the original code of a relatively simple indicator, when everyone is using the same source...OF COURSE the end results look much the same. Stop the self promotion and name calling.

[Harry]

Harry....when converting the original code of a relatively simple indicator, when everyone is using the same source...OF COURSE the end results look much the same. Stop the self promotion and name calling.

The way I have coded the indicator is rather unusual. You have not even bothered to change a single line of my code or the order of the variables ...

If you publish the indicator here in the download section, please refer to the website of Connors Research (www.tradingmarkets.com) and further refer to the source for the NinjaTrader code, as you have entirely copied it. That said, I am not against your publishing the indicator here. Referring to the source - that is Connors Research - would just be fair.

[photog53]

Ok....clearly I have taken this thread down a path that is neither helpful to the Ninja Support Forum, nor is it doing me or Harry any good. Harry aka "fattails" is an excellent coder...(true statement...no sarcasm intended). I have in fact learned a few tips and ideas from him.
The point of this thread was to get a piece of open domain software into the hands of a Ninja Trader member who needed it. The source code for this is fairly straight forward and I stand by my assertion that C# translations on simple code can/and do sometimes come out looking the same....in much the same way researchers sometimes do the same work in completely different locations. Instead of saying what is what I thought....which is that basically good C# code sometimes comes out the same....I took the lower road and snapped back a snarky reply. I apologize for my snarky tone....I am embarrassed that I let a comment cause me to contribute to sidetracking a helpful forum.

[billo]

Hello
Can you share an e-mail address of Fattails with me ? I need to have questions answered on the ANA TSI Indicator.
Thank You
Bill