adjusting adx trend "threshold" for length

T

tortoise

Hi - As I understand it, ADX set at the benchmark standard length of 14 is said to show a trend when it registers readings in excess of 25. Setting aside, for a moment, the question of whether or not such a threshold is worth diddly, does anyone know how to adjust this threshold, proportionately, for different lengths. A straightforward algebraic exercise -- e.g. adjusted threshold = (x/14)*25, where "x" is the length in question, doesn't seem to do the trick. I suspect the correct adjustment is a bit more subtle, but I haven't found my way clear to the answer. Any thoughts, anyone?

Thanks
 
As I understand it, someone told you that
  • Trending=yes when ADX(14) is greater than 25
and you believed them.

Now you want to know how to compute the function f(N) in this second statement
  • Trending=yes when ADX(N) is greater than f(N)

Why not go talk to the person who told you the first statement, namely f(14) = 25, and ask them for a table of f(N) for all N? Then you could just use the table directly, or fit a polynomial equation to the tabulated data.

Or you could talk to a lot of people and get a lot of tables, then average them together to produce a table of consensus opinion.
 
Quote from uncertain:

As I understand it, someone told you that
  • Trending=yes when ADX(14) is greater than 25
and you believed them.

Now you want to know how to compute the function f(N) in this second statement
  • Trending=yes when ADX(N) is greater than f(N)

Why not go talk to the person who told you the first statement, namely f(14) = 25, and ask them for a table of f(N) for all N? Then you could just use the table directly, or fit a polynomial equation to the tabulated data.

Or you could talk to a lot of people and get a lot of tables, then average them together to produce a table of consensus opinion.
More...

I'm continually amazed by the propensity of folks to reach conclusions without any supporting data. To wit...

As I understand it, someone told you that
  • Trending=yes when ADX(14) is greater than 25
and you believed them.

I don't know where you found this "understanding" but, I assure you, it wasn't to be found in my post. Feel free to read my words, precisely as written.

As for the "person who told [me] the first statement"[sic], I believe Welles Wilder established this threshold, although I may be mistaken here. In any event, I'm less-than-sanguine about the use of such arbitrary metrics, but that is beyond the scope and intent of my post.

Ok, back to the original question....would anyone with some knowledge/understanding of ADX care to shed some light on this? I suppose I'm unclear on the logic behind the commonly accepted trend "threshold" (which I do understand to be 25 with the default 14 period setting). Also, whether the logic of the 25-14 relationship can be adjusted for different length settings through a simple ratio -- e.g. threshold = (length/14)*25, assuming "25" is, in fact, Wilder's stated trend "threshold".

Thanks
 
Input: Length(14),NoTrnd(20),Trend(30),Ref(20);
Plot1(ADX(length),"ADX");
If Plot1 < Ref Then
SetPlotColor(1, Yellow)
Else If Plot1 >= Ref and Plot1 < Trend Then
SetPlotColor(1, magenta)
Else If Plot1 >= Trend Then
SetPlotColor(1, green);
{Plot2(Ref,"Ref");}

{ADX Alert Criteria}
If MRO(Plot1 < Plot1[1], 3, 1)[1] > -1 AND Plot1 < Plot1[1] Then
Alert("ADX has just turned down")
Else
If MRO(Plot1 > Plot1[1], 3, 1)[1] > -1 AND Plot1 > Plot1[1] Then
Alert("ADX has just turned up");

easily adjustable.
hope it's what you were asking
 
Trending = yes when ADX(N) is greater than f(N), please give me the formula for f(N)

BigBubba has supplied one answer: f(N) = ADX(N)[1], the prior bar's value of ADX(N).

I will supply another answer: f(N) = 25

BigBubba is saying that trending = yes when ADX(N) is rising. For all N.

I am saying that trending = yes when ADX(N) is greater than 25. For all N.

I feel confident you can find others with other opinions. Perhaps the prior suggestion, to average the opinions together and form a consensus, isn't a terrible idea.
 
You must log in or register to reply here.
Back
Top