Recently, I discovered Technical Analysis (TA), which is the science of anticipating market trends and movements based on technical indicators (price and volume charts, etc.), as opposed to fundamental analysis which tries to anticipate the market direction based on economic indicators.
I was initially quite skeptical. One assumes that people's
investment decisions and implicitly market movements are
completely random and thus impossible to predict. After reading a
few articles and a couple of books, I understood that TA makes
actually a lot of sense. In Technical Analysis of the
Financial Markets, John Murphy says: "There are three
premises on which the technical approach is based:
1. Market action discounts everything.
2. Prices move in trends.
3. History repeats itself."
There are aspects like the relatively large number of investors and the correlation that exists between their behavior (when the herd moves in one direction, it is difficult, psychologically, to move against it) that make statistics work (more or less) in this case. However, TA is by no means an exact science (or the market wouldn't work anymore the way it does now - I have to provide this disclaimer in case somebody will want to sue me {smile}.
Now, as a novice to TA, I would like to know what other people think and what their experience has been with TA. Does anybody know of any statistics on the subject: average gain obtained by means of technical analysts as opposed to more usual buy and hold type investment strategies, how many people use these techniques etc.
I am interested in TA not only from a practical standpoint but from a scientific one as well, in case other Colloquyans are wondering what the Stock Market has to do with intellectual inquiry.
MR: I like the book "A Random Walk Down Wall Street". According to academic studies that apply various technical schemes to past history, none of them work as well in the long run as simple buy and hold ... especially considering fees, although computer trading has decreased the cost. Technical analysis is much loved by stockbrokers, because it generate buys and sells, which is how they make their money.
AMi: Some observations: particles which have quasi-random behavior at quantum level (Schrodinger) combine to create very predictable macroscopic systems. It is thus possible that if the number of investors were of the same order of magnitude as the number of atoms and molecules in macroscopic objects, the stock market would be as predictable. Some TA methods are based on the Fibonacci numbers, other methods employ spectral analysis to derive market cycles.
WPP: Sounds to me like one of the basic ideas behind chaos theory. You might check to see if anyone has applied chaos theory to the stock market. I would be surprised if they hadn't!
MR: Assuming one of these esoteric technical methods worked, its effectiveness would soon be neutralized by everyone else catching on ... this is the "efficient market" at work. To succeed, you would need to continue coming up with new technical analyses that paid off. There are probably some people who can beat the S & P 500 over 10-20 years, but not many. And how do you know who they are upfront?
TsC: I'm also interested in technical analysis. Much of the market movements are due to the psychology of people. For example, what makes someone think Amazon.com at $220 (late April) was a good buy and $110 (now) is a must-sell? What makes the great change in the view of the future of the company while there're no important news in this period, no earning report, etc.? Psychology.
The basic TA such as advance/decline and support/resistance worked a hundred years ago and still work now. That means most of the money is still in the hands of people who are psychologically driven. TA promises to take advantage of the mass psychology. But it is not easy to perform well with TA because the players themselves are human. Some say you need both TA and a good execution, i.e., how much to risk and how to overcome one's own psychology.
WHK: Excellent, excellent, excellent! I had intended to post something just like this, but you made the salient point that others so often miss! - That is, Economics is, after all, a Social Science, subject to the whims of humans, individually and collectively, and as such is much more complicated than mere mathematical formulas, statistics, past history trends, etc. We still are a long way off from understanding, much less predicting human behavior. Thanks for the input!
TsC: I think I want to make my point clearer. Although we cannot predict the very detail or the whole picture of the stock market, part of it is predictable. It just means that for the pattern that you can predict well, you can make it right maybe 2/3 of the time and wrong 1/3 of the time. In terms of making money, that's enough. According to the author of "Trading for a Living", who is a psychologist and a trader, although the individual psychology is very complicated, the collective psychology is suprisingly predictable. I found that the basic indicators such as advanced/decline lines are amazingly accurate, but how can one use them to make money is kind of an art.
WHK: The difference between Psychology and Sociology! - I was thought to be "not-with-it" when, during my studies to earn a Master's degree in Psychology, I challenged the use of statistics as a valid tool for use in Psychology. It tells us nothing about the individual -- example: we can say that 37% of the citizens in a county are Republicans, but it doesn't give me a clue about what the case is with Mr. Smith. For the reason you cite, Statistics is a valid tool for Sociology.
We can study an individual and make predictions (extrapolations), but I will continue to question the practice. Once you study more than one, you are not really dealing with Psych anymore, but the study of groups. I often get the feeling that the psychological "establishment" is trying so desperately for Psychology to be taken seriously as a Science and therefore needs to be quantified, as if that would lend an air of authority. I think statistics just doesn't cut it for that purpose... Remember that as far as people are concerned, behavior always devolves to N=1.
TsC: Interesting point. Maybe statistics is not a good tool for psychology as you defined. But how would the knowledge accumulated in studying individuals be applied if it is so unique and cannot be generalized in some way?
WHK: Indeed, that's the issue I was raising with my profs! - Can it be generalizable? Should it be? We can make assumptions, guesses, and with the help of non-quantitative "tools" estimate a person's trustworthiness, loyalty, aggressiveness, sanity, etc., [using body language, evaluating tone of voice, past history, and so on] but we still don't have a foolproof means of stating with absolute certainty what a person might do under a specific set of circumstances. Sure, we might be right 99% of the time, and some of us are better at "reading" people than others, but there are always those aberrations which pop up and surprise us.
If the goal, I've heard this stated somewhere, of dealing with human behavior is prediction and control, the quantitative approach is still a long way off - again, mainly pertaining to the individual, while groups are relatively easier to figure out. Psychology vs. Sociology ...
Returning to the stock market, we can get pretty close to a predictive tool if looking at how a group of people likely will react, and that is all that is needed for the most part in seeing where the herds is likely to lead; but determining beforehand what Mr. Smith or Mr. Jones is going to do isn't so easy... and, again, not too important, MOST of the time, unless he's an owner of many shares or fund manager who makes a decision about lots of money that skews the rest of the situation. Interesting stuff!
The other problem is that with humans there are too many variables. B.F.Skinner's claim that human behavior is nothing more than Stimulus/Response makes it sound simplistic. Maybe if all the variables were taken into consideration that would be true. The number of variables, however, is so enormous that it becomes impossible. So, although certain things can be quantified, we can never be sure we've got the whole formula, let alone being positive that what we do have is correct in the first place. Which returns to my contention that Statistics is an inappropriate tool for Psychology, where N=1.
I'm rambling on here, but one last point ... it is even more difficult when dealing with such things as IQs. Intelligence is, after all, a human construct; and there is not even full agreement about what it is, so measuring it will remain problematical. Therein lies a whole 'nother thread of possible discussion.
JI: Statistics goes from the many to the one; i.e., from groups ofidentical particles to some general conclusion. Probability goes from the general (the many) to the specific based on the results from statistics. The result of probability is a statement that such and such an individual, in a given group of hairdressers (for example), has a probability of 80 percent of being female.
For example, we are given a group of identical particles. Let's say that we are only concerned about people who are hairdressers as the occupation. All other variables (height, weight, ethnicity, etc.) are ignored. Therefore, each particle is identical because they are all hairdressers.
We then use "gender" as a category for the random variable. We survey 4,900 hairdressers at random throughout the United States. We find that 4,200 of them are female, while 700 of them are male. Statistically, it is a valid conclusion to say that most hairdressers are female (5/6), while only a minority are male (1/6).
Now we go from the group to the individual. We are blindfolded and are presented 4,200 hairdressers in an auditorium. We are allowed to select one hairdresser in the auditorium.
For that individual, the odds are good (over 80%) that the hairdresser is female. If I were placing a bet, I would bet female rather than male in this case. This is how probability is used as a tool.
WHK: Certainly I understand what you've said (I know how it works; I helped write a book about Statistics and Probability for graduate Psych students and your explanation is fine), but I think you're missing my point. Using your example, you've still come up with a conclusion that 80% are female hairdressers, which tells me nothing about any given individual. I'm not talking about odds ... (you've addressed a GROUP of hairdressers). I'm questioning the aspect of certainty about an individual; it's not there; i.e., the tool doesn't match the subject.
JI: For this utility, I believe statistics and probability can still play a useful role in the "soft" sciences such as sociology and psychology. They give us only trends so that state, federal, and local governments can allocate resources properly.
WHK: I say, definitely "yes" as to Sociology, "no" as for Psychology ... that is, governments are concerned about groups (trends), but not about specific individuals.
JI: I understand your point. If an individual is nondeterministic, we can never completely successfully predict the behavior of the individual, or the fate of an individual with certainty.
However, if the individual has a certain past and has exhibited a certain trend, we can within certain errors predict the fate of the individual. A person with cancer, diabetes, AIDS, etc., depending on the condition, will have a shortened lifespan. Probability can predict this.
A person from a shattered home, abuse, lack of education, and surrounded by poverty and violence in the streets, with a low to low normal IQ, will probably wind up in prison before too long, or committing one or more crimes for which he will inevitably be caught. This is also borne out by probability.
WHK: That's exactly it! - and I have no argument with the rest of your post either. It's a matter of certainty vs. probability ... and the tools that are used to ascertain each.
JI: I believe that a democracy does respect the right of the individual and is concerned about the individual. For example, me, as I write this. Where will I be at this same time next week? Probably doing the same thing. N < 1. But I could also be struck by a crashing airplane and die. N = 1 (game over). We live in a massive probability cloud.
WHK: Good point, further making my case: you used the condition of 'probability' again. So given all sorts of statistics about you, we still can't predict with certainty whether you will die next week or not (I hope you don't!) - so the use of statistics in an individual case is not helpful.
TsC: Let's say two strangers enter an elevator. If they don't talk to each other and there is no intended interaction, by your definition, it sounds like this is a case about the psychology of two people. Now, suppose one say 'hi', there is interaction, then it becomes a sociology problem. Isn't the role played by the 'hi' too high?
WHK: I don't see where you're going with this. What does your reply have to do with the connection between psychology and statistics?
Instead of just saying 'hi', he may be at the beginning of an attempt to sell mutual funds! Does that keep the thread to stock market activity? {smile}
I hate to go banging on about this; I'm really not trying to argue the case against statistics, just that it's an appropriate tool for numbers of people but not for a single individual if we're seeking certainty in behavior.
TsC: I just want to point out that one cannot divide psychology and sociology simply by N=1 or N>1.
EM: Hello everyone, I do love discussing the stock market and economic systems in general. I don't think that the form of economic system means much to anyone except the owners. Whether the coal miner, factory worker, teacher, postal clerk, or the military, whatever, is employed in a capitalist or socialist system does not much change the nature of the work, or the rewards or opportunities. At the very top of whatever economic system in use, the rewards become very disproportionate, and nepotism is the custom.
Much of the American public lives by a host of myths, some of them very sad and forlorn. I visit the AOL political chat, and watch people arguing endlessly about the virtues of hard work, and hard earned dollars. One would think that the only dollars that exist are hard earned ones. The two richest women in the United States are the wife and daughter of Sam Walton. Tell me again about hard earned? Hard earned dollars are those earned by stoop labor and sweat shop female employees. None of them have E-Trade accounts, or make 40 thousand dollars with a phone call, or e-mail message.
To return to the AOL political chat groups. Many of the participants seem to be hard right wing politically, including the chat room hosts. Yet, when the discussion becomes relaxed and informal, at times it does, we see that the participants have as many money problems as any other ordinary group of working class people. Even more, if we then go to visit the business investor online chat groups, we do not find any of the people who are such devotees of the political chat forums. Which is to say, that the actual group of those people investing in stocks, or day trading, are not those who frequent the political chat groups.
Some years ago I looked into Mensa's Rich SIG, the requirements were by todays standards very modest, net worth $125K, exclusive of home and auto. What was of much interest to me at the time was that the money made by those people was primarily in real estate, 85%, the remainder by inheritance, stock market, or business ventures. There were also investment clubs, but I did not see that I.Q. had much of anything to do with the solvency of the participants.
JCC: A very fine point! {smile} While listening to the Tucson public radio, I heard the story of a local youth who thought it would be fun to purchase his own domain name a few years back. It amused him to register it presumptuously as WallStreet.com. He had no idea of the potential asset until a firm offered to buy his rights for a quarter million US dollars. He thought then to offer this domain name by auction instead and apparently received one million or more as the final price. That's a nice piece of work, interlaced with a small portion of luck and whimsy. Of course there's always that cosmic principle to contend with: easy come, easy go ... stated more generally in the simpler case of "easy go".
