Measuring trending behaviour

Belligerent_Drunk

Junior member
Messages
38
Likes
4
Dear all,

I was wondering myself something in the past week. I am afraid I can't explain it very clearly. In short, I want to know if there is a method (preferably statistical) to measure the tendency of price to "trend".

I mainly trade currencies, so I will use that as an example. I want to know if for example EUR/USD "behaves" better than AUD/USD on a given day. Does it trends more clearly with less deviations from the trend. Since imho the "cleaner" it trends, the easier it is to trade.

Of course, one of the issues is how to define the concept of trend.

I am curious of your ideas/suggestions.

Kind regards,

Koen
 
Indentity the trend market, some trader using method like as simple moving average with setting 200 to look for long term trending, if line sma show to down, its sign the market having tendencies to bearish and likewise, but small timeframe also mght occur small wave trend like as if there are news high impact can influenced currency instantly
 
More often than not the market trends against your position. I am stating the obvious because you know that already or your wouldn't be asking the question. So if you don't take a position, the market can't trend against you and will behave properly.

So the gist of it is that the trend is defined by positions.
 
good question ............many traders ignore the importance of comparing relative trend strengths of currency indexes and also pairs ...........for me it is very important and part of my edge ......why wouldn't you not be trading on the biggest trenders ?.......why waste time on the sluggish ones ?

I can suggest using my FXcorrelator regarding comparing index strengths / weaknesses ..........I don't offer anything here at T2win on Pair strengths....

N
 
I sometimes consult my "Trend Strength Indicator" to define but also compare trends and try to quantify them against each other. Its a very simple chart set-up -
HLC weekly bars against the 50EMA over a suitable period, say 6mths
count how many weekly H's, L's and C's are above the average
make this a %age
track the TSI score per week and watch out for the score changing

Other MAs and periods are available.
 
Dear all,

I was wondering myself something in the past week. I am afraid I can't explain it very clearly. In short, I want to know if there is a method (preferably statistical) to measure the tendency of price to "trend".

I mainly trade currencies, so I will use that as an example. I want to know if for example EUR/USD "behaves" better than AUD/USD on a given day. Does it trends more clearly with less deviations from the trend. Since imho the "cleaner" it trends, the easier it is to trade.

Of course, one of the issues is how to define the concept of trend.

I am curious of your ideas/suggestions.

Kind regards,

Koen

this is what I think you're saying...
if you were to trade something, you would rather trade something that trends better than another. Its not about the definition of the trend, and its not whether relatively one is stronger than another, but whether over a period of time something trends better.
for example on this chart below..
the chart represented by candles, over a period of time has shown a tendency to trend better than the OHLC chart
Have I understood you correctly? you don't care whether right now its stronger, or how strong its trend is now, but over a period of time it has shown a tendency to trend?
let me know if I have it correctly..
 

Attachments

  • trendingJPG.JPG
    trendingJPG.JPG
    112.9 KB · Views: 177
this is what I think you're saying...
if you were to trade something, you would rather trade something that trends better than another. Its not about the definition of the trend, and its not whether relatively one is stronger than another, but whether over a period of time something trends better.
for example on this chart below..
the chart represented by candles, over a period of time has shown a tendency to trend better than the OHLC chart
Have I understood you correctly? you don't care whether right now its stronger, or how strong its trend is now, but over a period of time it has shown a tendency to trend?
let me know if I have it correctly..

Hi Malaguti! (and the rest)

That's indeed what I mean. As a mroe concrete example, I've added two pictures:

First one, example of an almost perfect downtrend:
TEumzR7.png


Second one is a lot less "clean":
9hGwzAY.png


So the actual question is, is there a way to measure if some price trends "better". As NVP put it: why trade the sluggish ones?
 
Hi Malaguti! (and the rest)

That's indeed what I mean. As a mroe concrete example, I've added two pictures:

First one, example of an almost perfect downtrend:
TEumzR7.png


Second one is a lot less "clean":
9hGwzAY.png


So the actual question is, is there a way to measure if some price trends "better". As NVP put it: why trade the sluggish ones?

then I'm sure there are a few different ways of approaching the same thing.
personally, I'd try to create an indicator and your inputs could be as Tomorton suggested..the number of closes below an MA for example could give rise to a simple count function..a bit like the below would be the result

compared to the not so trending one

my inputs was actually the number of highs produced (given that its a higher high, higher low that defines a trend..for me at least) but the same could have been done with closes below a MA as Tom mentioned.

This would be the most straight forward..the higher the count, clearly the better the "trendability" as it were

the other way you could do it could be the slope of the linear regression of the price over a period of time..the higher the slope, again the trend..however I'm not sure you can get an actual measurement of it..statistical as you mentioned earlier. so its the least best option in my opinion assuming you dont want to be eyeballing every chart, hence the reason for a statistical method i assume

as for your specific example of a nice clean trend (id argue its not clean at all by the way) you could simply have a condition that todays close is less than yesterdays, is less than the day before and could easily probably be written as
C<C(1) and C(1)<C(2) etc

so you see there are probably a few ways..you might want to explore each of them and see which one works for you. there are loads of free charting software that allows you to create your own indicator/filter etc and took me a few minutes to write one so you could easily do the same im sure

hope this helps
 

Attachments

  • count.JPG
    count.JPG
    79.7 KB · Views: 169
  • Capture.JPG
    Capture.JPG
    70.3 KB · Views: 169
I think I'd be influenced by the pullbacks.

With the not so clean one the pullbacks are quite severe right from the off. Conversely, the clean one has minimal pullbacks - again right from the off.

The "right from the off" is important because you're looking to get into a decent trend as soon as possible. Not waiting until it's so obvious that it's getting near finishing or flattening.
 
I wonder is there any trend strength indicator available which doesn't lags?
 
I wonder is there any trend strength indicator available which doesn't lags?

I don't know for sure, maybe there are certain indicator that not included as lagging indicator, but might pssible giving false signal, I think if any trader thy can reading the chart pattern and behaviour, indicator is number three
 
then I'm sure there are a few different ways of approaching the same thing.
personally, I'd try to create an indicator and your inputs could be as Tomorton suggested..the number of closes below an MA for example could give rise to a simple count function..a bit like the below would be the result

compared to the not so trending one

my inputs was actually the number of highs produced (given that its a higher high, higher low that defines a trend..for me at least) but the same could have been done with closes below a MA as Tom mentioned.

This would be the most straight forward..the higher the count, clearly the better the "trendability" as it were

the other way you could do it could be the slope of the linear regression of the price over a period of time..the higher the slope, again the trend..however I'm not sure you can get an actual measurement of it..statistical as you mentioned earlier. so its the least best option in my opinion assuming you dont want to be eyeballing every chart, hence the reason for a statistical method i assume

as for your specific example of a nice clean trend (id argue its not clean at all by the way) you could simply have a condition that todays close is less than yesterdays, is less than the day before and could easily probably be written as
C<C(1) and C(1)<C(2) etc

so you see there are probably a few ways..you might want to explore each of them and see which one works for you. there are loads of free charting software that allows you to create your own indicator/filter etc and took me a few minutes to write one so you could easily do the same im sure

hope this helps

What are your specific rules to identify a down trend ?
 
It is my own and I am not prepared to share in details, but if you carefully read my posts you will get an idea.

I have looked through the last 6 pages of your thread and all I can see are channels. Whatever your semi-secret strategy is it cannot determine a trend strength on price alone. All a chart or price can tell you is what's happened in the past. It can't tell you when the trend is losing steam because it can't tell you why the trend exists in the first place.
 
I have looked through the last 6 pages of your thread and all I can see are channels. Whatever your semi-secret strategy is it cannot determine a trend strength on price alone. All a chart or price can tell you is what's happened in the past. It can't tell you when the trend is losing steam because it can't tell you why the trend exists in the first place.

You can not pinpoint the exact place when the trend is over. But following the Highes and the Lowes is not a bad idea.
 
Top