How do you determine trend direction?

Discussion in 'Technical Analysis' started by jpvil123, Aug 28, 2016.

  1. And also with you.

    When the logic which drives your anticipation is Boolean, there is only one possibility. But it can be (and has been) a sonamabitch to achieve that Boolean state.

    Benoit Mandelbrot authored a great book called "The Misbehavior of Markets". IMO, there is a single word describing the results of his experiments with market data which is of utility. Although he didn't understand what he found, he told the rest of us about it.
     
    #101     Sep 14, 2016
  2. True. To achieve that boolean state is hard but primordial as it removes hesitation, ambiguity, when the market is actually made up of contradictions, uncertainties & Co. One really has to remove the degrees of freedom in its interpretation. To Keep it Simple. We either have to add or disminish our exposure over time. Nothing more ... Nothing Less ...

    I've read Mandelbrot too. Got to know him thanks to Taleb. What single word are you thinking about ? Fractality ? The only thing I remember even though there's more to his book than that.
     
    #102     Sep 15, 2016
  3. Taleb's books were great. He is a must read for anyone serious about the markets.

    It would spoil the fun if I told you but what is there is worth the search. Fractality certainly sums up the book but then that's not surprising. This finding surprised Mandelbrot.
     
    #103     Sep 15, 2016
    Apophenia likes this.
  4. Here is a trace from a few days ago during and after rollover. It serves to illustrate the utility of a TL. I draw channels but you don't have to. The start of the downtrend is in the early AM of 9/8. One can, in real time draw left to right TL's and basically just sit until price bumps into the RTL. With futures, most retail traders will exit before market close (vertical yellow line) and reenter after the open. At about 8 PM (these times are PDT) you hit your first bump and can either hold, reverse or exit and reenter.

    As a consequence of the failed breakout the RTL can be fanned to give the very top RTL on the trace. Things meander along till the early AM of 9/9 when the trend accelerates down. Readjust your trendline to reflect this. It does this twice more as can be seen.

    At about 8 PM on 9/11 the niceness ends and the market moves laterally. Can you trade this using TL's? Yes but as they say you must be nimble or you will get slaughtered. A better option is to 'move' to the second fastest TL and sit until it is approached and then make another decision - hold, reverse or exit and reenter. The trend does reverse about 5 AM on 9/12 and the process repeats in the opposite direction. It is seen that a trend line is a money velocity line (deltaPrice/deltaTime) so you can be as greedy as you wish.

    There are other things on this trace having to do with volume, price groupings and bar action. Play with them as you wish.

    Finally the vertical red lines are when the fed-heads did their thing, Rosengren first and then Brainard. You can readily see that the market was already moving in the direction of their 'fallout'. Doe this always happen? Of course not but it happens not infrequently.

    So the point is that money can be made using the simple expedient of a TL. That's all.
     
    Last edited: Sep 16, 2016
    #104     Sep 15, 2016
  5. IMO, the only true trendline is the Sperandeo trend line. It is what the smart money uses. Trendlines have influence/are felt over multiple time frames and highly correlated (r2 > 0.95) instruments. That said, if you wish to use some other TL protocol, please do.
     
    #105     Sep 17, 2016
  6. Oh ! Actually I just remembered something that's also of paramount importance (Beside Fractality).

    The fact that there is evidence of memory - Autocorrelation - with volatility. And not so with Price.

    Wasn't that ? =P

    Oh !! And that Time ain't actually "constant" ! Transcendental insight too.
     
    Last edited: Sep 18, 2016
    #106     Sep 18, 2016
  7. The trend is your enemy
     
    #107     Sep 20, 2016
  8. %% On a one year candlechart; a 200 day moving average can help-but its NOT a prediction. Wisdom is profitable to direct.
     
    #108     Sep 28, 2016
  9. You flip a coin, head (Bull) & tail (Bear)
    Once flipped if you think the coin's crazy,
    Then you switch ... Otherwise you say "Okay"
     
    #109     Sep 28, 2016
  10. bone

    bone

    We use three different criteria in order to determine a possible emergent trend development. And by intent, only one of them actually samples OTR historical data points. The two other criteria are rules - based.
     
    #110     Sep 28, 2016