Comments on: Petra on Programming: A Unique Trend Indicator

By: Petra Volkova

Petra Volkova — Wed, 04 Jan 2023 08:47:02 +0000

var Spearman(vars Data,int Period)
{
int* Idx = sortIdx(Data,Period);
var Sum = 0;
int i,count;
for(i=0,count=Period-1; i < Period; i++,count–)
Sum += (count-Idx[i])*(count-Idx[i]);
return 1. – 6.*Sum/(Period*(Period*Period-1.));
}

If you have Zorro, you can find the source codes in the file “indicators.c”.

By: Kael

Kael — Wed, 04 Jan 2023 04:19:21 +0000

Hello Petra,

Can you include the lite-c code for spearman here ?

By: Petra Volkova

Petra Volkova — Sun, 16 Jan 2022 09:21:33 +0000

I am not this familiar with correlations, but when comparing it with the Zorro code for the Spearman and Pearson correlation functions, I see that you are probably right. The CTI uses the Pearson correlation.

By: Giorgio

Giorgio — Sun, 16 Jan 2022 08:35:52 +0000

In spite of what John Ehlers writes in the original article – Spearman correlation – the formula provided is the Pearson correlation formula. Pls check yourself. Spearman is robust to any distribution of the data, while Pearson implies normality. If you wish, you can try implement with the actual Spearman and see the difference (in terms of lead-lag or noise)

By: Arnis Lapsa

Arnis Lapsa — Sat, 30 Oct 2021 06:13:03 +0000

`CTI(Close, 14)` yields me about 10% (which isn’t bad)

sort of don’t like it / haven’t found proper application

By: Quantocracy's Daily Wrap for 04/17/2020 | Quantocracy

Quantocracy's Daily Wrap for 04/17/2020 | Quantocracy — Tue, 05 May 2020 05:46:33 +0000

[…] Petra on Programming: A Unique Trend Indicator [Financial Hacker] […]