Which Moving Averages (and Other Technical Indicators) Do the Big Players Use...

Discussion in 'Technical Analysis' started by cwb1014, Mar 6, 2010.

  1. If you can't see that 'testing' will NEVER be able to account for the infinite number of combinations and scenarios that can, not only arise in the markets, but can be used by the trader himself when using indicator(s), then I believe we will have to agree to disagree.



     
    #21     Mar 7, 2010
  2. The citation most common is AUG2000, JF; Lo, et al.

    What is pointed out almost immediately is the error of academics and the FI.

    Most traders agree with their (academics and FI) seminal standard that is clearly expressed and which represents exactly how this mis-application garners the exact consequences of the error.

    So far your commentary for years and years which is in vogue with this convention, leads exactly to the deadend you find that you occupy.

    Obviously, the only neccesary step for both academics and the FI to consider is correcting their error. What is required for them to do is to examine the record of those who do not make the error that they do.

    I speak from the orientation of an Awardee of a NSF "Program of the Year". To run the gauntlet at NSF to emerge as a recipient takes more than what the NSF expects as input for the use of their grants.

    The issue is NOT the presumed differences between TA and FE. It is the error found in the seminal standard.

    By not going down the road of the trip the citation represents, a trader simply puts himself on the correct path as a consequence of critical thinking.

    The correct path includes five considerations: motivation; reason; orientation; experience and performance.

    I look forward to your response; skip your usual litany, if you please.

    I have posted the citation and occurance the seminal error fouind in the citation. See if you can address why you continue justify and make this same error.
     
    #22     Mar 7, 2010
  3. I would like to enquire of the management as to whether this thread is going to be properly moderated or whether the usual BS moderation is going to prevail?

    TIA.
     
    #23     Mar 7, 2010
  4. I agree, why is Hershey still allowed in here? he is to trading, as compost is to grass clippings...
     
    #24     Mar 7, 2010
  5. The main error that the Academics and FI makes is that do research on TA and FE to determine its predictive value.

    One chart is presented in the citation. Of 77 bars there is a variation in the first 20 and the last 9 for ewither volume or price, respectively. you will also noice that the most frequent two pabe formation appears 17 times in the possible two bar price only formation; there is no connection to the probable volume counterpart for these appearances.

    If figure one relates to markets; then what happened to all their further tables? All are price only related.

    If trading were examined from a different basis than the one that is espoused, then entirely different and better perfomance would relate.

    As you see in the subsequent responses just above, there is no substance to t zones year after year commentary. And there is no minimum semblance of moderation with respect to any seroius consideration of any academic or FI papers in ET.
     
    #25     Mar 7, 2010
  6. "It seems to me that the answer to this question is this: those that the big players (e.g., hedge funds, institutional traders, etc.) are paying the most attention to. Because they are moving the most money, they have far and away the greatest impact on the markets when they act."

    Greatest players have no holly grail MAs. There is simply no perfect indicator in universe.

    As a beginner, I spent countless hours trying to find one.

    After about 10 years of trading I can make money with almost any common indicator or without.

    The only requirements for me are basic price charts with correct COHLV, right state of mind and trading capital that is not required for my family's urgent needs. As clarification, I do not consider myself a greatest trader. However, I managed above average return consistently for quite some time now. I used to be a failure in 2003-2004 market.

    As a preference, I use Slow Stochastic, RSI, MAs, FI, MACD with default settings. If someone would slightly modify my settings - it would hardly affect me at all. The indicators are like white dividing lines on the road. Nice to have but not required for driving.
     
    #26     Mar 7, 2010
  7. It is just unfortunate that there is no moderation in this thread. Were there moderation, there might be more diverse participation among those who are conversant regarding the current literature and research.

    The cited document goes down a less that useful path to reach a set of stated goals. Often, as a consequence , the citation is used to prove points by those who think the TA does not work. In fact, the presentation is an example of how not to approach the stated defined problem.

    In the case of this paper, the result turns out to be an effort that does not solve the problem the presenters continue to have.

    My past comments adddressed the beginning of the paper where the problem was set forth and where several statements were made in regard to how to scope and bound an analysis to be able to draw conclusions regarding the stated problem.

    Basically the paper presents the following work:

    A. Recognize TA

    B Comparing & Contrasting TA with FE (QF)

    C. Context of information for predicting the future

    D. Joint distribution of variables.

    E. Issue of noise (I)

    F. Automation of TA (II)

    G. Report results (III)

    H. Monte Carlo simulations (IV)

    I. Conclusion (V)

    My view is that the problem in trading is NOT predicting the future. Making the case for which is better, TA or Quantitative Finance, by analysis of market data to determine infomation content is not where the rubber meets to road for attaining performance of the application of capital.

    Acknowledging there is a joint distribution of variables and then interjecting a one size fits all mechanical constraint firmly and finally introduces an error that prevents any useful work result. This was an administative decision that simply destroyed any time spent.

    Noise may be a consideration of some tooling for doing analysis and determining results. What is the effect of dealing with noise at the expense of the value of a method, however? This paper makes the case of how erroneously dealing with noise destroys the work effort.

    Automating TA can follow one of two forks in the road. One can use data to find patterns or one can define patterns and search for their existance and productiveness. Another paper does the opposite of what this paper does and in doing so it explains what the third opportunity really is.

    The last three parts G, H and I have a value in that they scope and bound what using market information could mean for the process of making money. The focus is componding capital through prediction using a constrained orientation driven by pre research arbitrary considerations, a very costly orientation.

    Shredding research is a common enterprise that is not very valuable. By examining a failed attempt it is possible to create a Phoenix from the ashes. Who benefits from the effort? Anyone who takes advantage of the new opportunity, in part or in full.

    There is also the advantage of being able to examine the results of using a more specific and focussed research strategy and its products.
    I'll post a couple more times today and then come back another time in the future.
     
    #27     Mar 8, 2010
  8. Elsewhere, by example, I began some parallel considerations to the cited paper.

    One is a drill using two kinds of charts to prove that page 1729 which presents a volume strategy was an ill chosen concept.

    a greek symbol for volume was used. I will use v instead.

    there it was determinedto divide a pettern cycle into two equal parts and compare the volume of each part. from thsi arithmetic comparison the pattern could be applied under several categoies: increasing volume decreasing volume or neither. Neither comes about because the portions are contstrained arbitrarily.

    Thus quintiles are subdivided into three classes by volume. The quintiles are reloaded with samples every five years. 10 patterns are tested in each quintile.

    In five years each pattern is traded about 7 times. The hold time is fixed by the pattern's "l" a value arbitrarily set by the workers.

    the capital is compounded using an added factor of holding a fixed number of days after the end of the pattern.

    so capital in in markets and trading is going on and compounding is taking place.

    they doubled down on everything to assure that each in parallel set of tests were fairly close to one another i nthier outcome.

    Just by looking things over roughly, it is possible to understand that this type of trading would be called position trading since daily data is used , all patterns are 35 days long and the hold is a time out after 3 days following the completion of the pattern. Some how an entry and exit have occured in relation to a completed pattern.

    There is no overlap of patterns as a proviso.

    Before all of this is happening, noise is dealt with. It is averaged out first, then the 10 designed patterns are found and trading is compounded using 5 year periods over 31 years.

    Patterns involve points and in these patterns there are limits for the range of values for different relative points in the patterns. Say a couple of tops are involved. the contstraint is that the tops be within 0.75% of their average. Another pattern comapares two values using 1.5% of average as the outside difference in value. For many patterns, these constraints are chosen.
     
    #28     Mar 8, 2010
  9. Could you please provide examples of this research?
     
    #29     Mar 8, 2010
  10. It was brought up that in other research on TA the subject was usually the performance of rule based systems. This research did not use any rules, they said. They focussed instead on TA as defined by the occurance of patterns that could be used to predict. And there was a comparison to the QF of FE to find out which worked best.


    I looked at an alternative research affort done by Michael Harris (December, 2006). He used a search engine to find patterns the markets create. the pattern is then expressed in trems of the componet extremes of the pattern (only inequalities are used). Thus the broadening pattern of Lo does not appear in Harris, since it is not found through searches for sequences of events in markets using logical search strategies.

    This is like setting up sights on weaponry where one set of sights is to one side of the target and the other set of sights is to the opposite side of the target. Few targets that are present are hit by either set up for research. Both leave out rule based TA trading research.

    Rule based TA trading is studied ad nauseum it turns out. Some punch up 5,000 individual rules and test their individual performance rate. See Marshall, et al, 2009 for example.

    To polish it all off why not do a study of studies. See The Profitability of Technical Analysis: A Review.

    ET has gone through all of these one way or another.

    This means that the world does know what the big players use and how well what they use works. This is all moot.

    What is the take way for a person or business that wants to optimize their trading using TA?

    Several things come to light.

    1. Learning from other's mistakes is very important. And it is a time saver.

    2. There are big chunks of great value right out in front of all traders who will go into a modus of critical thinking.

    I have stacks of three ring binders full of researcher's results. The two biggest pieces that stick out head and shoulders above the others are the mistakes of Lo and the compass that Harris allows a person to discover. First the compass. Harris shows how patterns can be described by bar groups. In the simplest set of groups he incisively delivers the punch line: one two bar pattern. Lo looks at five years and finds 7 trades a year average; they are fixed in discovery duration and decription and are time out exit based.

    The results:

    All patterns begin with SYM's roughly speaking.

    Do not be prescriptive of what you are looking for because that is all you will find.

    3. All trading works based on TA, no matter what it is named. What defines TA is the order of events in markets (Harris, 2006) and the joint distribution of the variables (Lo, page 1707).

    4. The rules studies back up the CW toatally. CW has no Holy Grail nor does big money use any semblance of TA or QF that is successful. discovering its absence is a load and clear FI statement.

    5. To trade sucessfully using TA and indicators is an individual or group process achieved over time. The details, I beleive, anyone can discover through critical thinking. Either the learning or the discovery process has to be done uins critical thinking. As a person goes through his process, it is very important to not make mistakes and accept the mistakes as beliefs. It comes down todealing with the joint distribution and how the order of events is dictated by the joint distribution.


    It is possible that serendipity could happen and a trader would be seemingly born instead of being made. By realting to sucessful traders for 50 years, it turns out that doing great trading doesn't require much market time and harvesting the offer is just a routine mental activity like any other. The payoff in compounding is the last third so getting the basic build up of capital should be done in a couple of years as a reasonable goal.

    The define DO NOT DO of the markets is predetermining patterns and looking for them.. NEVER use predetermined values for anything. NEVER do time out type trading.

    For most people, they will NEVER think in deductive terms. To learn to trade as an EXPERT NEVER use inductive reasonning.

    If it is at all possible keep an open mind and get AWAY from the following as soon as possible: up, down opposites; entry/exit trading; predicting; betting. What can take the place of each of these respectively, is: looking at price move dominantly to the left and non dominantly to the right all relative to trend line; focus on holding during trend continuation and changing sides of the market when the sentiment changes (stay on the right side of the market, always); Knowing that you know replaces predicitng (in trading there is only NOW and you have to know that you know in NOW); knowing that you know replaces betting (the market only has two sides; there is NO NEED to bet on one or the other; the right side is always available from the tells of the market).

    So what is the BEST signal from the market? It IS the most frequent signal: the HOLD signal. You can look at the window that ends the HOLD; it has two sides, left and right. You can keep drilling down to finer and finer panes to see the end of HOLD kick in. You have plenty of time to find the price where the turn happens; it is stated way ahead of time and price just comes to that limit as a pattern completed.

    I get to go away for a quite a while to do recovery. THINK and build your mind by doing drills.
     
    #30     Mar 8, 2010