Separating randomness

Discussion in 'Technical Analysis' started by 931, Jul 21, 2020.

  1. 931

    931

    Recommend materials on separating randomness.

    Could be books,videos etc.. not even directly related to markets or algos.
    For instance DSP lectures, lectures about noise filtering algos, some aspect of math in general or whatever gave you useful clues.
    Whatever you found to be useful or holding clues about separating useful info form noise aka "finding order in chaos".

    I found https://www.3blue1brown.com YT channel extremely useful for math.
    Custom coding&effort put into making visualizations in order to make complex math simpler to understand.
    For instance FFT gets broken down using simple visualizations. Before i found it hard to understand on wikipedia or by some lectures.
     
    Last edited: Jul 21, 2020
    .sigma and murray t turtle like this.
  2. %%
    200day moving average;
    +50 week ma= Greatt Visual.....................................................................
     
  3. .sigma

    .sigma

    I follow that guy on tweet ah, he’s pretty entertaining and smart.

    as far as separating randomness, how would you accomplish this via technical analysis?
     
  4. 931

    931

    Developing and using algos/models to find useful info in chaotic data classifies as TA right?

    IMO it also means separating randomness/noise to some extent.

    Separating randomness might be just as hard as finding useful info...
     
    Last edited: Jul 21, 2020
    .sigma likes this.
  5. .sigma

    .sigma

    ohhhhhhkay gotcha I know what ya mean now.

    extracting the alpha from the beta

    the signal within the noise

    the calm within the chaos

    it’s out there man, probably hiding in plain sight.

    I think a lot of it comes down to the interpretation of data, having the right measurements, looking at the right metrics, reading it rapidly, reactionary immediacy, and order execution.

    for example, I am a big fan of visual finance. When I can see a visual representation (like a chart) it helps a lot. But I’m also a quant guy, in the sense that I like quantified data from the chart.

    for example, let’s say I’m looking at a 6 month daily chart of TSLA. Visually it’s making higher highs and lows, booming up. Thats good to know. But I could also count how many days were up compared to down in that 6 months. I could also measure the daily range and divide it by the total iterates and get a feel for the average daily move etc etc
     
  6. 931

    931

    Yes.

    If we would have all beta noise, then alpha could be viewable without hiding?
     
    Last edited: Jul 21, 2020
  7. Real Money

    Real Money

    Probability Theory really is very good. But a lot of it is too technical and too impractical for trading.

    Knowing how taking products and ratios of random variables will affect the distributions, or how linear combinations affect the variance based on independence and correlation can give you great insight into risk management. But all of the various techniques for parameter estimation get so complicated that only quants and statisticians can understand it.

    It's the ideas that are important, not the technical mastery.

    An example would be the classic 60/40 split. Owning the sum of inversely correlated assets reduces risk, while still achieving the benefit of returns. Different weightings can add return, but will affect the portfolio risk.

    That idea is being applied to every imaginable combination of assets, on all timeframes, and with every kind of trading objective.

    It makes for very interesting markets.
     
    Last edited: Jul 21, 2020
    931, fractalize and userque like this.
  8. 931

    931

    Is there improbability theory also to approach from other side?

    It makes sense.
    Some of those people can work for years on single problem without reaching breakthrough.

    But failures are also very valuable info.

    Because while getting new ideas all your failures will be consulting you.

    Noise might be valuable if it's separatable.
     
    Last edited: Jul 22, 2020
  9. Grantx

    Grantx

    Defeatist attitude. It is easy to understand if you put the effort in.

    Also, a certain level of technical proficiency is very important. Become self sufficient with your application of stats don't dismiss it because it looks hard.
     
  10. Real Money

    Real Money

    If you think estimation theory is easy then you don't understand what I meant.
    A certain level, yes. But knowing markets and trading is much more important than mastering Ph.D level statistics.
     
    #10     Jul 22, 2020