Re: [amibroker] Williams %R

[Peter Carr <pcarr@xxxx>]

On Mon, 5 Mar 2001 at 12:32:02, Tom Presley <tpresley@xxxx>
wrote about [amibroker] Williams %R:


Peter Carr posted an AFL for the Williams %R indicator a few months ago.
Thanks Pete, this gave me a helpful starting point.


Glad you found it useful.


After comparing the charts to those from www.stockcharts.com I found 
Petes formula didn't match. I updated the formula as show below which 
does match:

Thanks for pointing out the discrepancies. I also have to give thanks 
to Richard Cloonan who (via Email a while ago) also pointed out that the 
formula he was familiar with used the 'high' rather than the 'close'. 
This is practically identical to your posting other than the range/scale 
adjustments.

I hope Richard, you don't mind me taking the liberty of reposting your 
modifications here for list members to consider in this discussion.

==== Richard Cloonan's Williams %R Mods ====
days = 20;
num = hhv( high, days ) - close;
den = hhv( high, days ) - llv(low, days );
rate = 100 - (num/den * 100);
graph0 = rate;
graph1 = ema( rate, 10 );
=====================================

I find Richard's mod version a useful addition to my collection of 
custom indicators, and your version will undoubtedly be included among 
the ever growing AFL folder :-)

One point that I would like to address, particularly for newcomers to 
Technical Analysis and this list, is that I have found that there are 
often different alternatives, variations and uses of an indicator, from 
writer to writer. These may even differ from those of the originator of 
the indicator as analysts experiment and test them to develop their own 
(possibly) 'unique' system that 'works' for them.

You may indeed consider 'my' (Larry Williams') formula to now be 'fixed' 
in terms of your own perspective and the 'authority' of the website you 
quoted. Nevertheless, I am still content to consider it an alternative 
given the status of the books (and authors) I used to develop it and the 
following quote from one of them on which I based my formula:

"...Today's close is subtracted from the price high of the range for a 
given number of days and that difference is divided by the total range 
for the same period." Technical Analysis of the Financial Markets by 
John J Murphy 1999

In one of the figures in his book, Murphy shows a 20 day %R with 
overbought at => 80 and oversold at =< 20.

I've also found the 'closing price' version to be reasonably effective.

Is your negative scaling as per Larry Williams? I have not as yet taken 
a peek at any of his published work. I have to admit difficulty in 
interpreting your chart - though this is purely due to my exposure to 
20/80 or 30/70 overbought/sold thresholds.


Note the line where rate is calculated. Williams %R is supposed to range
from 0 to -100.


Could you identify your source for this scale please? I have not read 
widely on this indicator, but those book I have read show 0 to 100 and 
often point out that it is reversed to that of Stochastics.


However Tomasz program won't put indicator lines at -20 and -80 which 
is the normal crossover points for this indicator, so I adjusted the 
output to 0 to 100.

There's seems to be various accounts in published material of what is 
'normal' for this indicator. CMIIW, but also there appears to be some 
tendency to standardise indicators (oscillators?) to the 0 to +100, 
20/80 or 30/70 scale.

I have to say that AFL is very flexible and certainly allows for tweaks 
and modifications to any formula. Customer friendly :-)


Peter
--
======================================================================
Peter Carr
Email: pcarr@xxxx ICQ: 22586379
======================================================================