Trendiness of Days - assigning a value

Discussion in 'Technical Analysis' started by TGregg, Nov 20, 2003.

  1. well, one possible use of an intra-day "trendiness" indicator could be to gauge inter-day momentum.

    This is completely speculative and theoretical (and I have done zero work on this), but intra-day trendiness could have a relationship to inter-day momentum.

    I recall some articles mentioning how the market seems to "buzz" at tops - calm interday patterns but a great deal of intra-day volatility below the surface.

    Again - I have no evidence for this, but it could be an interesting (and useful, for a change) approach to investigate the fractal nature of markets.
     
    #21     Nov 22, 2003
  2. ig0r

    ig0r

    <a href=http://mathworld.wolfram.com/CorrelationCoefficient.html>Calculating Correlation Coefficient</a>

    <a href=http://www.mega.nu:8080/ampp/rummel/uc.htm>A more advanced approach</a> This one has some interesting stuff that could be applied to trend analysis
     
    #22     Nov 22, 2003
  3. Man, no way anyone can make money using this shit
     
    #23     Nov 23, 2003
  4. DK_

    DK_

    HERE'S HOW YOU DO IT:

    run a repeat for each tick, finding the ticks which are either the highest highs or lowest lows for the 20 (or so) bars on either side of the one being examined. The number of such bars found is the number of trends in the day + 1.

    IE:
    OPEN AT LOW, CLOSE AT HIGH DAY would have 2 such bars at open and close, therefore 1 trend

    OPEN AT LOW, PEAK HIGH MIDDDAY, CLOSE AT LOW DAY would have 3 such bars, one at the open, one midday and one at the close, 3 = 2 trends.

    just tweak the 20 value for whatever time increments you use and voila
     
    #24     Nov 23, 2003
  5. This is very knowable for anyone. You make the key point for trading in the last paragraph. Another person here mentions Range trading. All three of these aspects are continually knowable so the trading focus may be maintained almost all of the time.

    Once people get off the "market only does three things" mentalitiy, they can move to the consideration of how to detect the market operating point at all times.

    This thread has many many posts about what happened after the fact. What is most important is what is possible to anticipate. There are many kinds of trending segments, many kinds of "chop" and many possible "ranges" to consider. Having a set of measures as to which is possible at any time really puts the icing on the cake.

    You can see which traders in ET are aware of these possible delineations.
     
    #25     Nov 24, 2003
  6. Today was sort of another double trend day...........

    Response to second para.:

    It might be possible to find the True Range for successive market days over a long span of time, then find the average of all those measurements. Any day where the True Range is greater than average is a Trend Day and vice versa, every day where the range is less than average is a Range Day.

    Response to third para.:

    You could also measure within a day as Acrary did...just find the distance between the High and the Low. If the close is equidistant between the two call that a Low Trend Day. Depending upon how close the Close is to either the High or the Low determines how much of a trending day has occured. You can measure this using a percentage measure away from the midpoint of the day's range...the further away from the midpoint the greater the percentage the Close has moved resulting in a trend measurable by percent.

    I enjoyed reading the posts and seeing you guys' contributions.
    :cool:
     
    #26     Nov 26, 2003
  7. ==============================
    [1]Like hiking or walking ,like to measure it lots of ways.

    [2] Previous day closing price to closing price, the day you mentioned, regular market hours;
    & Market makers edge - measure it from daily open price to closing price.:cool:

    ==
    Love learning--Solomon,trader king
     
    #27     Nov 29, 2003
  8. 2 cents:

    Perhaps ADX, which was originally designed by Wilder to measure the trendiness of any bar, in terms of its directional movement, in comparison to its surrounding bars during an observed period of time, could be exactly what you want in a practical sense.

    :confused:
     
    #28     Nov 29, 2003
  9. You could also use moving averages. If they point in one direction and are steep it's trending. Also if price hits the moving averages and bouncing, it's also trending. When the moving average lines are all lined up and flat, it's not trending. This is a simple method without having to use to man indicators, lines etc. Without indicators you could just see if it keeps making higher highs and higher lows etc. This means the market is trying to break resistance and keeps bouncing off support. (Harder to use interday though because of the whipsaws.)
     
    #29     Nov 30, 2003
  10. privateisland,

    Very interesting your "could" and "simple method".

    Please "could you" express your "simple method" in a little mathematical formula. Then we will talk a bit further.

    nononsense
     
    #30     Nov 30, 2003