Can linear regression analysis really predict the future?

Discussion in 'Strategy Building' started by tradrejoe, Nov 4, 2009.

  1. #121     Nov 12, 2009
  2. Maybe we should rehash the question at hand in order to widen the discussion and avoid "fighting" for a YES/NO answer?

    S'thing like this:

    "Can linear regression analysis consistently predict the direction of prices (up-flat-down) within a "short" (tbd) timeframe with a probability better than tossing a coin?"

    or similar...:confused:
     
    #122     Nov 12, 2009
  3. I have always known of Maestro to be a person of integrity on this forum. After a lot of thought with the very limited understanding of splines I would say yes since they are adaptive.
     
    #123     Nov 12, 2009
  4. MAESTRO

    MAESTRO

    Absolutely! Our discussion here is not about "HOW" we spot the patterns but rather "WHY" those patterns exist. The underlying premise of my approach is probabilities, randomness, stochastic, clustering of events and spontaneous synchronization. The arguments from others here are: the markets are deterministic because all the participants behave rationally and act to their advantage. The main disagreement here is that I believe that no matter what actions market participants are taking they all sum up in a very much random process in the price/time space. Following this train of thought I also argue that the traditional TA (head and shoulders, moving averages RSI etc.) do not have any reason to work as they presume that people “MUST” act certain way when they see those patterns. My approach is that nothing can work on a single run. The only way to exploit the markets is to implement law of Large Numbers where stable distribution patterns could be played out to our advantage. The arguments from some people here are also based on their personal “success” story, however, I believe that any personal trading history is a type of a random walk that could well drift into positive territory and prompt those who happened to be “Lucky Random Walkers” to believe that their success is somehow a result of their wisdom, intuition, experience, or proper implementation of the TA tools. That should put some more “oil” into the fire of this discussion! :)
     
    #124     Nov 12, 2009
  5. MAESTRO

    MAESTRO

    This material is way too complex. The easiest way to play with splines is through MATLAB. They have prebuilt libraries of splines that you can apply to any data set.
     
    #125     Nov 12, 2009
  6. MAESTRO

    MAESTRO

    In one of my previous posts I have already presented the answer to your question. If you missed it I will post it again, but do not miss it again please :) Mean Deviation of m random walks with n steps each = sum(squareroot(2n/pi))/m; This is the ultimate answer and the main building block for any trading strategy! Enjoy!
     
    #126     Nov 12, 2009
  7. Yes
    OLS is a proven Mathematically derived estiamtion..assuming Linearity (Not the case)
    also a relative indictor
    check TA Using three month
    will do good for a month IMO

    again check tol for the same period 3 months...southbound
    http://www.prophet.net/analyze/javacharts.jsp?symbol=ta
     
    #127     Nov 12, 2009
  8. jem

    jem

    Reasonable summary but I take issue with the loaded statement - personal success story.

    I ask again if a person produced evidence of thousands or trades with yearly returns of 300 percent or more can you really call that a result due to randomness.

    What if they traded at a 65% clip winning trades with losses smaller than wins.

    In short how does one prove the markets are not random to a random walk theorist?
     
    #128     Nov 12, 2009
  9. I don't think this should be shocking at all, although I would question that reversing the strategy is really anything more than random. The only reason MA cross systems are popular is because they are easy to pull off with our shitty retail software. I would also imagine that if your testing on 15 minute bars, considering the range any one 15 minute bar could have, half of what matters in the test is how any one bar randomly sets up as the cross happens. To not control for this to me makes the test totally meaningless.

    The ultimate problem though IMO with any system that is using some kind of average is you are not judging the magnitude of potential wins and loses..If a regression even had a 99% chance of being right, it wouldn't mean much if it had a 1% chance of a 100% loss. You would still go broke in the long run.

    As far as market randomness, to me the market is clearly not a stochastic process but there is enough noise created based off wrong information to make stochastic modelling useful. We know the market is not made up of a million people flipping coins at various times to buy or sell. More importantly IMO, even if it was there is not enough liquidity for everyone to change directions exactly as they would like to. If a fake news story came out tomarrow that the NYSE blew up, the DOW would not go from 10k, to 500, back to 10k in an hour..there is not enough liquidity for that which to me is exactly what makes trading possible.
     
    #129     Nov 12, 2009
  10. MAESTRO

    MAESTRO

    I respect your opinion. Seriously, I do. I respect anybody's opinion as the matter of fact; that is what makes this discussion an interesting one. I have met many mathematicians who reject non-deterministic theories for any of the human activities. However, in order to stand by my beliefs I always use majority as an example. Yes, one individual can be successful. If you flip a coin 10,000 times and randomly bet on each outcome there is a good chance that you will end up with the profit. It, however, does not prove anything. On another hand, if you take 10,000 people that flip the coin 10,000 times each the average of their performance will be close to "0". If you apply this mental experiment to the trading public you will realize that the "Average Trader" breaks even at best (minus commissions, of course). That is the only true experiment that sheds some light on the randomness of the markets.

    Having said all of that I encourage you to keep on thinking about this phenomenon even if you still believe in your own theory. I like very much when people have an opinion and fight for it. I respect that!
     
    #130     Nov 12, 2009