Talk:Technical analysis/Archive 2

Archive 1

Archive 2

Archive 3

Moving average

Perhaps someone here can more clearly explain the content on Moving average (finance) about exponentially weighted moving averages. As the article states, an EMA should theoretically assign a nonzero weight to all past values of the stock price and gives one definition of EMA. I believe the second definition is an attempt to define what is meant by N-day exponentially weighted moving average — a term I have heard, but wasn't immediately sure on how it would be defined. The article implies that the two definitions are alternate but basically equivalent. Are price values more than N time periods in the past excluded in practice but not in theory? btm 08:36, 10 January 2006 (UTC)

I believe the short answer is no, that in practice it's normal to use more than N days. The first N days is about 86.5% of the total weight, so you want to go past there, to maybe 99% or 99.99% or whatever you feel like.

That 86.5% incidentally is actually 1-1/e^2. I suspect there's some mathematical significance to cooking the "f" factor to come out like that, but I don't have any reference on it. -- Kevin Ryde 22:26, 11 January 2006 (UTC)

Perry Kaufman's "Trading Systems and Methods" gives a very thorough and mathematical explanation of various moving averages (simple, exponential, geometric, pivot-point, triangular). If someone has questions on moving averages, the book is a very good reference. Mind you, Perry's book is pretty intense. co94 Jan 13, 2006

To be precise, the exponential moving average (EMA) and the so-called weighted moving average (WMA) are quite different. The weights in the weighted moving average have constant differences between adjacent weights (eg 1/2, 1/3, 1/6 which provides a 3-period WMA) and the weights in a exponential moving average have constant ratios (eg 1/2, 1/4, 1/8, ...). Strictly speaking, the weights for the EMA should go back indefinitely, but in practice it is ok to curtail the calculation when the weights become very small. Incidentally, I think the term weighted moving average is an unfortunate one, as it sounds too general, but it's now entrenched Elroch 12:43, 15 February 2006 (UTC)

EMH

In academic community, even those proponents of EMH no longer thinks stock price follows a random walk. This is for sure. —Preceding unsigned comment added by Yiyu Shen (talk • contribs)

Anonymous deletions

Unfortunately this page has been subjected to what could be described as vandalism by anonymous "contributers" - large scale deletions of material without justification. For example, a reference to the important historical fact that Japanese candlestick methods were used hundreds of years ago was deleted. Technical analysis is a field where some people have very strong, blinkered views. Wikipedia is not the place to attempt to impose those views on others. This article should be a balanced, NPOV introduction to a broad, and very varied field. If anyone disagrees with this, they should at least have the courtesy to say so using a registered handle. Elroch 13:47, 15 February 2006 (UTC)

Feb 18 changes

I made a few changes to the part about fundamental analysis that I think make it a little easier to understand. Also, is this article best served with introducing fundamental analysis in the second paragraph? Is that necessary? co94

March changes

I clarified the part about Buffett and Lynch because it seemed that introducing them as opponents of technical analysis is a bit biased because they are exceedingly well known and very successful without using technical analysis. I thought it was best to mention that they are fundamental analysts and as such are often in disagreement with technical analysis.

Also, I removed the portion about economists for two reasons. First, as someone else wrote earlier in the article, technical analysis and fundamental analysis are the two major schools of thought regarding stock market investing/trading. Economists are not in the business of stock picking, investing, etc. Secondly, I could make a reasonable case that economists actually do use principles of technical analysis. For instance, US Weekly Jobless Claims are a notoriously fickle number. Most economists do not look at weekly changes in the reading as much as they look at its 4 week moving average, a technical analysis concept. Moreover, economists look at trends all the time. For instance, as of March 2000, it is safe to say we are in a rising interest rate environment. I.e., interest rates are trending upward. Moreover, any economist would agree that the economy is cyclical. It moves from expansion to deceleration to recession to recovery back to expansion. History tends to repeat itself. Sounds like a technical analysis concept! co94

I'm a physicist, an experimental physicist, with only a recent interest in stock-markets. I found the article useful [as at March 14th], and informative. Analysis of empirical trends and patterns in data is certainly science, and covers a broad range of techniques. Of course it is crucial to ascertain whether the pattern found is statistically-significant. Without actually specifying a very specific methodology or approach I don't see that TA as a whole can be proved or disproved. Different TA indicators could give contrary predictions for the future, but since the market is non-deterministic, indicators will only lead to a certain probability of a given outcome, based on the data (and time-period) used to formulate the hypothesis. For a particular TA strategy to work, all that is required is for the past characteristic of the market to continue to apply over the duration of your future investment, and not be dwarfed by other statistical fluctuations, or a deeper shift of fundamentals. It seems perfectly reasonable to me that a hybrid of TA and a general awareness of fundamentals is a good approach. Given that you will be trying to return above-market returns in a timeframe shorter than your historical analysis, I guess a large part of the 'art' is choosing which TA indicators will be most likely to apply significantly over the foreseen investment period... which sounds like a gut-instinct based to some degree on ones interpretation of fundamentals! 57.66.65.38 17:55, 14 March 2006 (UTC) Andrew - www.techmind.org

Fair enough, Andrew. Truthfully, what everyone is trying to do is predict the future and no matter what tools are employed, at the end of the day it is only a "best guess". Dennis Gartman, a legendary trader, boils down trading to 'watching the technicals and knowing the fundamentals" He might think the fundamentals support buying a stock, for example, but if the trend is downward, he might say stay away until the trend reverses. If say, everyone was buying gold mining companies, yet the price of gold was plummeting, then he would say the trend is intact, but the fundamentals completely disagree and again stay away. At the end of the day, there are zillions of different ways to approach the market and no one will ever find the Holy Grail of predicting the future.

With regards to your last statement, there is one thing I want to point out. Remember, that most technical indicators will have both entry and exit strategies. A sound technical strategy will not only tell you when to enter a trade but also when to get out. Often exit signals are triggered only to be followed by entry signals at a later date. This is something that another poster pointed out about how TA strategies can be inefficient. All those buys and sells generate commissions and eat away at returns. co94

Techniques of technical analysis

Frankly, the article at the moment, and perhaps particularly the "theory" section, is over-simplistic and suggests that the techniques of technical analysis are a lot narrower that they really are. I personally know many people who make a living from technical analysis, and am familiar with the approaches of many others that I have not personally met. It is true that the identification of trends is a common factor to many systems, but contrarian approaches (virtually ignored by the article at present) are used successfully by a substantial number of traders. One popular category of methods identifies when a market appears to be ranging and looks for trades back into the range when the price is near the boundaries - this is ignored by the article at present. I personally know a small number of successful traders who are happy to look for trades directly against a trend near to an extreme (or even in some cases at an extreme) when they believe that the trend is about to fail (don't try this at home, guys :-). Dow theory may be historically significant, but it is hardly the mainstay of 21st century trading, which is based more on time series analysis using computers. For example regression analysis techniques that can take into account higher derivatives than the first (which might be associated with the trend) can better predict the likely movement of the market, and will sometimes predict reversals, while falling almost entirely outside the scope of the article.

Hence, would people agree there is a need for a considerable broadening of the scope of this article? Elroch 23:26, 6 April 2006 (UTC)

Verifiable, authoritative content

The quotes from Warren Buffet and Peter Lynch are amusing (deliberately) and, although rather insubstantial, deserve a place in this article. In addition, there is a place in the article for verifiable, reasoned, unemotional statements about technical analysis by its opponents, but there is a lack of such statements at present.

However, I have removed the following three statements as examples of those that are not in compliance with wikipedia guidance. If this was an article about Manchester United football team, there would be little place in it for quotes like "some critics think they are a bunch of wazzacks who can't play football for toffee" and the standards in this article should be as high. Specifically the three statements do not comply with WP:V in several ways.

• though purists will argue that they can't be combined - a very odd statement that is contradicted by tbe fact that they have been combined, (in fact by some very well-known successful speculators). Who are these purists, anyway, and why would we want to include a statement by them that is patently false?

• ...and by its critics "is much derided as hocus pocus" - vague, unsubstantiated statement about the vague opinions of some unspecified critics.

• Critics argue that its claims to offer insights into investor psychology are absurd, and that it has less rational basis than astrology or the study of UFOs."

In the last two cases, which are based on the reference [1], we have examples of quoting a vague statement by one person about the opinions of some other unspecified persons, which is not adequate for an encyclopedia article. Who are these anonymous critics and why should we include second hand, emotional paraphrases of their opinions in an article about technical analysis? If such opinions are to be included, it is absolutely necessary to identify the specific people who have stated their opinions, and to put these opinions in a separate section (or preferably a separate article) unless they have some factual content about the subject of the article.

Incidentally, it is interesting that the article referred to mentions other material that is far more relevant. It refers to the work of Richard Olsen, a highly respected academic who is a world authority on time series analysis, which I believe he has used successfully for years. Isn't this the sort of topic that would improve this article? Elroch 16:20, 7 April 2006 (UTC)

from his website: "Dean LeBaron, founder of Batterymarch Financial Management in 1969, directed the firm's pioneering advances in the mid-1970s in the application of computer technology and modeling techniques, first in the US market and then in international and emerging markets. Today, Batterymarch is one of the investment management subsidiaries of Legg, Mason, which manages over $250 billion through several independently operating firms."

"...Dean pursues his interest in complexity through Complexity Digest [www.ComDig.com], a webzine he publishes, and through association with the Santa Fe Institute and the New England Complex Systems Institute and their linking of complex adaptive systems to dynamic social systems, including investments. ... he has written The Ultimate Investor and The Ultimate Book of Investment Quotations, published in Spring 1999 by Capstone Ltd. In 2002, Mao, Marx & the Market, Treasury of Investment Wisdom, and Book of Investment Quotations were published by John Wiley & Sons. His web site ... has 50,000 hits per week."

"Dean LeBaron received BA and MBA degrees from Harvard University, was a Baker Scholar at Harvard Business School, and holds a chartered financial analyst designation (CFA)."

In short, I think he has enough authority to state the obvious ... that a lot of people consider technical analysis to be "hocus pocus." And I do think that it's important that a new reader coming to this article knows that a lot of people consider this stuff pure bunkum. There are some arguements for TA, there are many against it. Let's cite sources here. I'll start in a bit with a few results from google.

searching google for "tea leaves" and "technical analysis" give 505 hits,

searching google for "hocus pocus" and "technical analysis" give 1,010 hits,

searching google for "pure bunkum" and "technical analysis" gives only 4 hits,

picking one just cause it said PBS, gives the following quote [2] "One of the main reasons technical analysis is scorned by so many is that its proponents can't tell you why it should work. That's why it's so often referred to as voodoo or black magic. Its practitioners claim to have discovered mysterious forces that govern the markets, but exactly what those forces are or why they work - that's still unknown after all these years. Human events, war, peace, revolution, technological breakthroughs, economic policy - these aren't what really determine securities prices. They are mere foam on the waves of…of…of the mysterious forces. Technical analysts hate being compared to astrologers, but you can see why the comparison is hard to avoid." from Goeff Colvin of Fortune Magazine and Wall Street Week."

now please let's not pretend that TA is a well respected, accepted theory, and just explain what it is, and provide some documentation.

I am again removing the phrase ", though purists will argue that they can't be combined" as provably false opinions of unspecified "purists" (whatever they might be) are not appropriate for a factual article. As iron-clad justification for this, I refer to the chapter on Randy McKay in Jack Schwager's "The new market wizards". Randy McKay grew an investment of a few thousand dollars into several tens of millions using a combination of technical analysis and fundamental analysis. I have many more examples, if you believe he may have just had a bit of luck. If this phrase is replaced a third time (as I hope will now be wisely avoided), I will follow the wikipedia procedure for resolution of disputes WP:DR, as it is unacceptable to include material that doesn't have factual content about the topic of the article (and which, in this case, is directly misleading).

Incidentally, I was surprised at what a tiny number of hits your amusing web searches came up with, most of which most are pages that promote the use of TA. A more relevant, specific search for "trading" AND "profit" and "regression analysis" came up with over three hundred times more hits than the sum of your three searches, many of which are studies that show that certain technical analysis methods provide an acceptable level of profit. Elroch 19:10, 7 April 2006 (UTC)

Reality check I'll leave out the "can't be combined" phrase for now until I have a good source for it, but please don't accuse me of editing in bad faith. You've remove that phrase (and some other things) 3 times today (14:11, 11:20, and 17:27 (6th)) Smallbones 19:36, 7 April 2006 (UTC)

Actually the only thing I have removed three times is the one falsehood, I have left other drivel for now, but it is not appropriate to be left indefinitely. I'm not sure that I have successfully got my point across that it doesn't matter if the president of Latvia says "some critics are of the opinion that the moon is made of green cheese", it would not make it an appropriate statement to put in a factual article about lunar geology.

This article is about technical analysis - i.e. mathematical predictive modelling of markets, and not about unsubstantiated opinions about technical analysis. The scope of the article is what technical analysis is, how it is used, research on how effective it can be, and anything else that has genuine substance. Certainly there is room for academic studies that show that some methods aren't effective (although this is entirely obvious, and the fact that some methods have been shown to be effective is more substantial). Technical analysis is like weather modelling: using good models in either provides a better estimate of future probabilities than a random guess, but neither is easy, and neither provides certainty. Elroch 00:26, 8 April 2006 (UTC)

What the Heck Happened to this Article?

It is a bit distressing that academic wonks continously and repeatedly (almost in cycles) come to this article on technical analysis to challenge the basic premises of technical analysis. Moreover, we seem to be re-inventing the wheel here as this article has been revised a million times before and I just dont get why.

There are so many incorrect statements in this "revised" article that I dont know where to begin. First off, I have never ever ever seen technical analysis defined as "the use of numerical series to predict trends." Where is this reference?

Technical analysis does NOT reject the Efficient Market Hypothesis. Efficient Market Hypothesis does not state that future price movements are a Brownian Motion (Random Walk Theory does.)

Paragraph 5 of the article is completely POV.

The elements of Dow Theory have been completely removed. This is the cornerstone of technical analysis.

"Technical Analyis is not based on any standard theory in economics or finance"?????????????????????????????????????? Who the heck wrote this? TA is all about supply and demand.

I could go on. It is obvious to me that, as was the case about a year ago, that academics have come in and dont want to describe technical analysis so much as attack it.

PLEASE READ PREVIOUS COMMENTS BEFORE YOU MAKE CHANGES TO THIS ARTICLE. THE CHANGES MADE ARE ALMOST ALL INCORRECT. THE PERSONS OR PEOPLE WHO MADE THE CHANGES DO NOT UNDERSTAND WHAT TA IS AND ARE MORE CRITICS THAN ANYTHING ELSE IT SEEMS.

Sorry to get all in a huff, but it seems like I went through all this last year and now the same. I am going to reinstate the original article. If people want to make changes then discuss them first and please dont include something that is based only on what your professors say or what you read in one article on the internet. Admittedly, the article should have been referenced better and I will get to that. Still, the person that rewrote much of this article really took a lot of the factuality out of it. Please be fair to all and discuss things before making massive changes. co94 April 17 2006

I (as someone who depends on TA to make a living) agree with you about the awful editing of this article. I feel I shouldn't bother wasting any more time making constructive edits when some narrow-minded ignoramous is going to erase anything that doesn't agree with his personal opinion. However, in reply to your comments (1) I get the impression it's more people with a narrow exclusive view of technical analysis who are the worst editors of this article; (2) This may not be a definition you are familiar with, but it is a good one that covers methods that don't explicitly use charts, as well as all methods that make decisions based on charts (3) I have a standard text on economics by my side and have know economists who trade for years, and it seems clear that it is not the fact that prices react to supply and demand that makes TA valid, it is the fact that supply and demand (and hence prices) change in a way with some sort of identifiable pattern to it; (3) TA does reject even the weak form of the efficient market hypothesis, which says that there are no methods, fundamental or technical, that achieve excess returns over the normal average return from the market. The semi-strong form of the EMH says the market is a machine that perfectly and instantly discounts all information in price (ROFLMHO). Anyone with any common sense realises that the market, and the sum total of its participants are as much like a perfect machine with perfect discounting of information as elections in a banana republic; (4) I can't see anything to disagree with in paragraph 5, as long as you take the word "trend" in its most general meaning (as used in statistics) i.e. a trader needs to have an anticipation that prices are likely to move in a directional way (as opposed to a random walk); (5) If you read the article, you would realise that Dow theory (which is certainly has an important place in the history of technical analysis, but which might not be even considered by the designer of some modern systems based on behavioural finance models) is still well-represented in the article. 80.0.184.11 23:27, 17 April 2006 (UTC)

Responding to 80.0.184.11's points

On 1) Perhaps. I feel that much of the "incorrect" revisions are by people who are not users of TA or have narrow views. Last year, as you will read above, I had to go tete-a-tete with academics who repeatedly attempted to discredit TA as opposed to describing it. Even before I revised this article a year ago, the article on TA basically derided the entire practive as nonsense. And much of the revisions recently appear to me to, again, not so much as describe TA but to discredit it.

On 2) Unsure. I think that definition a bit overstates the basic premises of TA. We are walking a slippery slope when the opening statement claims TA is a form of numerical series analysis but later says TA is more art than science.

On 3) Disagree in principle. EMH is very close to the premise in TA that, "markets discount everything." EMH says that the market discounts new information instantaneously so exploitation is impossible. TA takes that argument one step further. It says that information not yet revealed to the market is embodied in a security's price. So I dont see how TA is at odds with EMH, which, mind you, is an increasingly discredited theory. And yes, I agree with you, EMH ignores the reality of the marketplace. In fact, according to Murphy, "..academics have very eloquently stated (via EMH) the need for closely monitoring price action..." (Murphy p21)

On 4) My point re paragraph 5 is that by listing an obviously anti-TA statement implicitly comparing TA to UFO sightings is inherently POV. A reader might stop reading about TA right then and there becaue the implication is not that TA doesnt work but it is the realm of crackpots and bozos. I am not going to include in the article statements that lambaste fundamental analysts, (there are plenty out there) because it would skew a reader's opinion. Dont want to advertise TA.

On 5) Perhaps I misread the article but I didnt see Dow Theory mentioned in the article. I only saw it removed from the beginning of the article. Again, Dow Theory is a cornerstone of TA. It should be prominently featured in any article on TA so I was unhappy to see it completely removed from the beginning of the article.

Overall, I felt the original article gave a fair explanation of TA criticism and I was very careful not to claim that TA could deliver guaranteed results or "Easy money" or whatever. As we all know, the burden of a wikipedia article is not "truthfulness" but "verifiability". The academic community cannot agree on whether TA is viable or not. Some think it is; others think it isnt. Some market participants love it. Others dont. That's all there is to it.

As I have stated before, I note with raised eyebrow that the article on "fundamental analysis" is painfully short. Perhaps the people who went to town on this TA article should improve that opposing article instead of erroneously modifying this one?

That being said, I rewrote a chunk of the article and started to include references. I am sure this is not the cleanest way of referencing an article but I wanted to get some in there. I will continue to input references as I have the time. co94 April 23 2006

This article is certainly better now than the last time I had an argument with someone who wanted to include 3rd hand quotes with no factual content trashing technical analysis. I like the comment suggesting opponents of TA go and write a decent article on fundamental analysis. There are too many people who waste their time trashing articles on topics they don't agree with rather than writing articles about things they think are worthwhile. One comment I dislike is one about academics proving TA doesn't work. Anyone who made such a broad conclusion would lose all credibility, as it would require essentially proving that each market is a random walk. Reputable negative studies prove much weaker results which might indicate that a particular class of methods were not effective in those markets examined.

I cannot agree with the comments on the relationship of TA to the EMH. The EMH (which is empirically 100% discredited from being more than an approximation with significant inaccuracies) takes as axioms things which logically imply that TA cannot make an excess profit (however the point that the market is "mostly random" is an important one to bear in mind while trying to pin down the part of the market movement that isn't random). As the EMH axioms are provably false in real markets this should not concern technicians unduly. The very idea that there is a right price for the market is a ludicrous fairy tale made up by academics - even they have found it impossible to find any theory which tells them how anyone might determine inarguably what the right price is. Of course we technicians realise that the market has a price which at every instant accurately reflects the sum total of the opinions of its participants but that this price changes without any information beyond the price movement itself driving it. A good example is when a market is rangebound - a devotee of the EHM could only, to be true to his beliefs, say that there was news hitting the market each time it reached the edge of the range and sending it back the other way; a technician with his feet still on planet Earth might say price tested the low and having tested it be likely to test the high, until the range failed (with more than a false break). Anyhow, great to see people making positive contributions to the article. Elroch 20:49, 23 April 2006 (UTC)

May 2006 Changes

Modification claiming that behavioral finance regards technical analysis as voodoo deleted. It was unreferenced, biased, and had a ton of grammatical errors. In fact, "Behavioral Finance" by Goldberg and von Nitsch, ISBN 0-471-49784-3 shows how behavioral finance supports some of the premises of technical analysis. co94 May 2, 2006

Deleted sentence, "Ironically, behavioral finance....consequence of heuristic biases." Sorry, not trying to be daft here, but I had no idea what that sentence meant. It just seemed very out of place. Can you explain more clearly what is meant by the aforementioned and from what source it comes? co94 May 5, 2006

Included the bit about Buffet and Dow Theory because I thought it was interesting. The quote is verbatim as reported in an analyst report from BNP Paribas published May 10, 2006 by Sylvian Brunet. title: "Metals and Mining" co94 May 11, 2006

Reverted back to original description on Random Walk Theory. The revised description is not one that I think is correct. I dont think any of the other contributors know of that description either. Also, it didnt seem to make sense because in the first revised paragraph it stated that the market takes into account all information available in past prices while in the subsequent paragraph it said that future prices are independent of past prices. co94 May 22, 2006

You're right about the inconsistency. The statement that "future prices are independent of past prices." (which was in the text before my edit) is obviously incorrect and I've corrected it to "movements in prices are independent of past prices." As regards random walks and weak-form EMH, I've corrected the text to say that the random walk hypothesis may be derived from the weak-form EMH which is obvious (there's the minor technical difference that a random walk is required to have a stationary distribution of shocks, which isn't imposed in EMH, but apart from that the two are saying the same thing). Google "weak-form EMH"+"random walk" and you'll get plenty of instances on this.

Deleted blurb about EMH as a benchmark for testing trading rules. I think it is irrelevant for an article on technical analysis and better suited for an article on EMH. co94 May 25, 2006

Original research?

I deleted this section from the introduction as it seems out of place

A mathematical proof of the existence of technical analysis helps explain how fundamental and technical analysis work together. The well-known equation M = P/E, multiple equals price divided by earnings, while true is miscast. It states Multiple is the dependent variable, dependent on judgements made by investors over what the E (earnings, dividends, products, etc.) is and what the actual price is. M is the arena of the technician, the psych, the supply/demand situation. But the value of M is a judgement, as is E. Price is a given, you can look it up in the newspaper.

Therefore, price is the dependent variable, dependent on what we believe E to be and how much (M) we care to pay for it. In bullish times people pay excessive amounts (M), in bearish times, low M's. You cannot look up visions of E or M in the newspaper. Therefore we recast the same formula to become P = E x M (multiply both sides by E). Price is now shown to be the dependent variable, as it actually is.

Now one can see how fundamentals and technicals are dependent on each other. At times, fundamentals are more important to investors, at others the technicals, and every variation in between. But it is certain that unless P goes to zero, M must exist. Therefore technical analysis (M) must exist and must be examined to get the full picture of a stock market investment. Sometimes it has more weight than at others (tops and bottoms).

Technicians have many tools to examine each of these two areas, none 100% accurate, but it would be foolhardy NOT to look. What if ALL the technical indicators were bearish? Bullish?

Artbristol 23:44, 28 June 2006 (UTC)

You also deleted the neural net section without an explanation, so I'm inserting it back in. (It certainly isn't original research as you can see by the references) --Denoir 09:55, 30 June 2006 (UTC).

Sorry, I misread the revision history. --Denoir 10:08, 30 June 2006 (UTC)

June 2006 Changes

I reverted to older version because the introduction was completely and incorrectly changed. Art, the above explanation regarding the "mathematics" of TA is pretty much completely subjective and unverifiable. Whoever wrote the stuff about "psych" and etc in the intro looks way off. The section was riddled with spelling errors, exclamation points, opinions, etc. You cant make such massive changes to the article without discussing it first. Seems like most people had agreed on the intro to the article as it was. I've studied a lot of TA in the past. I've never come across much of what was in written in the revised version. co94 June 29, 2006

SMALLBONES, will you please stop implementing your agenda on discrediting both Dow Theory and Technical Analysis. I deleted the blurb regarding John Allen Paulos that claimed technical analysis is a pseudoscience. We had a consensus here about not using the term pseudoscience with regards to TA. John Allen Paulos also thinks that counting calories is pointless because it is impossible to determine the precise number of calories in, say, two different apples. So I assume you will have a lot of work to do on the dieting article. In short, Paulos' comment is misplaced here.

This article gives a good and fair explanation of criticism of technical analysis. There is no need to interject John Allen Paulos' opinion. If that's the case, I will enter a lot of traders and money managers who will say that TA works for them and this article will get very cluttered. There are a lot of people who say TA works. There are a lot of people who say that it doesnt.

Please understand that you do not own this page. It is a fact that many people, including John Allen Paulos, consider technical analysis to be a pseudoscience. The academic studies of TA are about 95% against. This stuff is not a minor footnote to this article, but needs to be upfront. You are violating the NPOV rule in this article by eliminating almost all evidence against TA. Smallbones 13:08, 1 July 2006 (UTC)

Smallbones, you are totally wrong. There is plenty of material in this article that criticizes TA. In fact, it is clearly labelled "Criticism of TA". The opening section comes right out and says that there are many critics of TA and that there are many studies of TA that conclude it has, "little, if any predictive power." Read the article.

With that in mind, have you even READ the Paulos link you sumbitted (or the book for that matter)? Smallbones, he doent call just technical analysis a pseudo-science, he calls FUNDAMENTAL analysis a pseudo-science as well. He is a Random Walker!!!!!!

Again, Mr. Paulos is contemptuous of pseudo-science, arguing that the superficially dispassionate ethos of fundamentals "does not make them immune :::::to emotional and cognitive distortion. The tango of exuberance and despair can and does affect estimates of stock’s fundamental value."

The most factual statement should read something like, "John Allen Paulos believes that technical analyis, as well as any other market analyisis or trading strategy, is a pseudo-science." By selectively editing out Paulos' disdain for all types of analysis, you imply that he only rejects TA. That's untrue and it is biased and POV. Paulos is a Random Walker. Random Walkers do not believe in TA or fundamental analysis. That is stated clearly in the article.

You have an agenda Smallbones, that is why I undo your changes. I take it you didnt go and edit the fundammental analysis article and include John Allen Paulos' comment or the article on EMH. You only want to discredit TA. As I have said earlier, a consensus was reached NOT to call TA a pseudo-science because the opening line of the article states that TA is non-scientific. It doesnt follow a scientific method; that claim is not made. It is more, "art than science." This article is loaded with cricism of TA. FAR more than the article on fundamental analysis. There is no way you can argue that I strip out all criticism of TA. I only want to ensure that the criticism is appropriate.

I made an appropriate and NPOV change to the TA Random Walk Section. co94 July 1, 2006

"Non scientific"

Saying that TA is non-scientific does not make any sense. It is perfectly falsifiable, which is the basic element of the scientific method. You define a hypothesis, build a model and test it. There are tons of published academic articles on TA models and results.

Generally speaking, we are dealing with a branch of applied mathematics that focuses on non-linear markov processes. Looking at a chart and making wild guesses is not TA, as some here seem to think. So called "Charting" may be a non-scientific method (due to the lack of formalization of hypothesis and model), but TA does not equal charting. TA is the the use of any type of system that tries to model signals based on historic trading data - or to simply find patterns in that data.

Serious TA which you will find in published articles, in more advanced trading software etc is perfectly measurable. We're talking various forms of time series predictions, transition probability analysis, classification etc

This is especially true for models that use adaptive systems (such as neural nets) as they adapt through performance criteria (hypothesis testing is the core) and statistical measurements are what dictate how the model is adapted.

So, the people putting "non-scientific" in the intro sentence either have a way too narrow (and incorrect) view of what TA is or don't know what "scientific" means. --Denoir 00:39, 2 July 2006 (UTC)

What you write above may be very true, but the fact remains that TA does have a lot of chart reading and pattern guessing. It isnt all nearly as mathematical as you describe above. I think saying TA is "mathematical" is a bit of an overreach. Saying there is a a flag pattern on a chart is certainly part of TA but hardly mathematical. I think the original definition is more suitable. co94 July 8, 2006

Well, that's charting, and I have no problem with saying that charting is non-scientific because it certainly is. But to say that TA as a whole is non-scientific wouldn't be correct as there are many methods that are measurable, falsifiable etc... --Denoir 12:44, 9 July 2006 (UTC)

TA & Charting

One of the major problems with this article is that on many places TA is equated with charting, which is dead wrong. Technical Analysis in general is defined by the axiomatic definition that using historical data you can predict future values of financial time series. This covers a very wide range of approaches, many that use a scientific approach. Traditional charting does not, but it doesn't mean that other methods are as vague. The article should probably differentiate between the methods that comply with the scientific method and those that rely on vague human interpretation of signals. --Denoir 18:08, 5 July 2006 (UTC)

Major problem? The introduction says that TA is primarily concerned with price charts which is very true. If you know of a different definition, please cite it. There are definitely more complex forms of TA like you describe which are more highly evolved, but they all try to predict market moves based purely on perceived patterns, etc. John Murphy's book on TA is widely considered the basic go-to book on TA and his definition is the one that was originally included. --co94 July 8, 2006

A web definition search gives a number of definition that don't necessarily refer to charting:

"A method of evaluating future security prices and market directions based on statistical analysis of variables such as trading volume, price changes, etc., to identify patterns"

"An approach to forecasting commodity prices which examines patterns of price change, rates of change, and changes in volume of trading and open interest, without regard to underlying fundamental market factors."

"A form of market analysis that studies demand and supply for securities and commodities based on trading volume and price studies. Using charts and modeling techniques, technicians attempt to identify price trends in a market."

Anyway, I do agree that charting is probably the most commonly used method, but to generalize and to assume (as this article does for the most part) that TA equals charting is very unfair. It's like imply that Astronomy equals Astrology because more people read horoscopes than they study positions of stars for scientific purposes.

Ideally, the article should cover the common assumptions of TA (i.e that there are trading patterns in historical data that can be used to predict future trading patterns) and then branch off in to describing the different methods. --Denoir 13:11, 9 July 2006 (UTC)

Statistical claims

I've removed a few sentences that make statistical claims that are not born out by the references. A paper only makes a statistical claim if it states something about statistical significance. This would usually be done using a fairly standard test statistic such as a t, F, or Chi-squared statistic, with the significance level sometimes summarized with a p-value. R-squared is not a test statistic, but rather shows "goodness-of-fit" which (formally at least) may not be that important.

The 3 papers sighted (still in the references) were said to make statistical claims on the following pages (after I removed them the first time)

• Lawrence p. 17-19 Nothing close to a statistical claim

• Birgel p. 6-7 Plays around with R-squared and a related "mean relative percentage error" but no significance levels or test stats given.

• Zekic p. 7-8 says something vague about "correctness"

In short, if you don't know anything about statistics, please do not make statistical claims in the artical. Smallbones 09:46, 11 July 2006 (UTC)

I absolutely agree with you that you should not make statistical claims if you don't know anything about statistics. I just wish you would have followed that principle before making unnecessary edits.

First of all you are confusing statistical significance with significance level. The first is a qualitative description of that a result is not just a random outcome from a given distribution or an sampling error. The latter is a specific quantitative measure of statistical signficance and there was no mention of it in the article.

Second, I think the problem is that you probably don't know much about neural networks. So let me enlighten you. They (at least the types used in the articles) are MSE and MSPE based and have some desirable statistical properties such as that you can read the confidence directly from the MSE (de-normalized). There's also a simple link to statistical significance.

You have the error criterion (MSE) which is

$\operatorname {MSE} (T)=\operatorname {E} ((T-\theta )^{2}),$ and you have consequently the SNR/PSNR:

$PSNR=20\cdot \log _{10}\left({\frac {{\mathit {MAX}}_{I}}{\sqrt {\mathit {MSE}}}}\right)$

And finally the SNR or PSNR relation to statistical significance is trivial and you can find it in the statistical significance article. The confidence that a result is not by random chance (i.e statistically significant):

$\mathrm {confidence} =SNR\times {\sqrt {\mathrm {sample\ size} }}.$

The steps above are a bit simplified, but I hope you get the idea. There are a number of elegant proofs of the connection between the outputs of sigmoid based feed forward backpropagation networks and statistical confidence of the results. It however involves stochastic calculus of variations applied to non-linear Markov chains and given that your statistics seems to be at the level of elementary hypothesis testing, it might be a bit of overkill in this discussion.

Now while I do agree that the three articles cited are somewhat vague in the result department from a statistical point of view, they do still implicitly confirm statistical significance. It's however not difficult to find articles with explicitly stated significance levels, like this one for instance:

SW Kim, AV Mollick, K Nam, "Another Look at Long-Horizon Stock Returns: Evidence from the G-7 Markets or Lin et al, Can the Neuro Fuzzy Model Predict Stock Indexes Better than its Rivals?" or perhaps even better yet this one which deals with the EMH as well and has a section deveoted to hypothesis testing: Skabar, Cloete Neural Networks, Financial Trading and the Efficient Markets Hypothesis

What I'll do is that I'll return the text but change the references to papers that explicitly state significance levels so that the claim is more easily verifiable for people with limited knowledge of neural nets and statistics. Ok? --Denoir 11:29, 11 July 2006 (UTC)

I'm going to look over the paper by Sacket cited at statistical significance. That characterization in terms of signal-to-noise ratio clearly has some surface plausibility, but is certainly not second nature to me. Michael Hardy 20:17, 5 August 2006 (UTC)

July 2006 Changes

I am unsure whether "Rule Based Trading" should be described where it is in the article. It doesnt appear applicable under the "Beliefs" section. Moreover, it seems more of a description of a particular area of technical analysis. thoughts? co94 July 16, 2006

Three-Dimensional Technical Analysis

Am I the only one feeling that this is a promotion for a book? AFAIK, this "3D TA" term is introduced by a book published one month ago. Does it already deserve to be mention twice in the history section? Regardless if it is good or not, I do not beleive that this book and its new "3D TA" term has pass the test of time (never heard of it before). In my humble opinion, it is not yet notable. mfortier November 10, 2006 Update: The (3D TA) section is now removed.

The same old story:

"Three-dimensional technical analysis" involves the analysis of price movements over time in multiple stocks simultaneously. It is thus a descendant of Dow Theory, as it focuses on the aggregate or relative, rather than the individual, movements of stocks.

In three dimensional technical analysis, traders need to consider not only price and time in a single stock, but they should expand their focus to similar information in multiple stocks simultaneously."

Better read Jesse Livermore's book published in 1940 and books about him: 1, 2, 3, 4, 5 Vugluskr 21:36, 30 March 2007 (UTC)

Thanks for the info. I am fine with adding back a reference to 3D TA, as long it it not done for book promotion (like it was before). Mfortier 23:39, 31 March 2007 (UTC)

Moving average - significance of lag

I'm only a beginner with this topic, so please excuse any misconception or odd terminology... but it seems to me that a vital feature of the moving average analysis is the lag. The article says if a share closes below the 200-day average then some will decide the run is over. It's important to know that the 200-day average lags the non-averaged data by 100 days (for uniform weighting)... so if a share closes below it, it actually closes below what the trend says the share should have been doing 100 days ago. This is a lot more significant than simply closing below the value predicted by the 200-day linear trend, which probably happens one day in two. As the article stands, a non-expert may misinterpret it saying that the latter is taken as a trigger to sell.--Russell E 00:33, 14 December 2006 (UTC)

Nothing is predicted, nothing is protected. Insiders will buy their stocks back near the year moving average also known as EMA51Wk. Or maybe not :-) Vugluskr 04:52, 3 February 2007 (UTC)

External link

Perhaps this external link may be useful Tracking the Trends in Technical Analysis, an interview with the founder and first president of the Market Technicians Association, Bob Farrell. Some insight into how the field has developed. Have Gun, Will Travel 00:54, 11 January 2007 (UTC)

History repeating itself

I have deleted this from the introduction because I don't think it adds anything:

on the basis that history repeats itself more often than not

As a "basis" this doesn't work, because it's just a restatement of the central hypothesis - a useful explanation of what is meant by "history repeats itself more often than not" would probably end up as a full explanation of technical analysis. In other words, it is dangerously close to circular reasoning: "technical analysis works because technical analysis works".

I don't think anything's really lost by leaving the sentence as "Technical analysis assumes that non-random price patterns and trends exist in markets, and that these patterns can be identified and exploited."

Mswake 15:29, 5 March 2007 (UTC)

Price × Volume indicators

I spent some days trying to understand why price and volume are multiplied in such indicators:

• Money Flow
• On-balance Volume
• Price and Volume Trend
• Accumulation/distribution index

I answered myself: "Maybe, it is better to divide them? Why not divided?" Finally, Dimensional analysis gave an answer to my question.

Please, check my contributions in these articles, especially language.

Vugluskr 07:31, 23 March 2007 (UTC)

Greenspan

I have removed the quote Greenspan quote because, despite the superficial connection between his reference to the behavior of prices and technical analysis, he is not asserting, as technicians do, that the future direction of price movements can be predicted from past price movements, but rather that prices fluctuate, and will continue to do so. --Benna 22:16, 3 April 2007 (UTC)

I couldn't disagree more with this. Greenspan, the Fed, and other Central Banks regularly use technical analysis. They are interested in market expectations, sentiment and key price levels. When I wrote a commentary during my time at a major sell side dealer, Central Banks were amongst the most common readers of my work. He states that prices behave now as they did in the past. He also states that human nature ties the future to the past. In other speeches, he has directly discussed how these patterns are repetitive. Sorry, but Mr. Greenspan's words belong. Sposer 21:43, 4 April 2007 (UTC)

It’s completely clear that Greenspan is not a technical analyst- see below for his own words, as reported by the Fed. You think that some of his ideas support the concepts of TA – but that is original research on your part, and doesn’t belong in Wikipedia. See WP:OR. To get Greenspan in here as supporting TA, you need to find him saying (in some reputable source) “I support TA” or “I use head & shoulders all the time.”Smallbones 16:15, 5 April 2007 (UTC)

Remarks by Chairman Alan Greenspan At the European Banking Congress 2004, Frankfurt, Germany November 19, 2004

From Federal Reserve

“The inability to anticipate changes in supply and demand for a currency is at the root of the statistically robust finding that forecasting exchange rates has a success rate no better than that of forecasting the outcome of a coin toss.2”

Footnote 2. The exceptions to this conclusion are those few cases of successful speculation in which governments have tried and failed to support a particular exchange rate. Nonetheless, despite extensive efforts on the part of analysts, to my knowledge, no model projecting directional movements in exchange rates is significantly superior to tossing a coin. I am aware that, of the thousands who try, some are quite successful. So are winners of coin-tossing contests. The seeming ability of a number of banking organizations to make consistent profits from foreign exchange trading likely derives not from their insight into exchange rate determination but from the revenues they derive from making markets.

It's not "original research" to say that Greenspan refers to market behaviors that technical analysts recognize; it's a statement of fact that is relevant to this article. What's more, the Fed not only uses technical analysis, it publishes technical analysis. This should be (and now is) mentioned in the article. Rgfolsom 17:18, 5 April 2007 (UTC)

The Boston Fed explicitly states: "The Stock Market Report is in no way an endorsement of any one mode of study or source of advice on which one should base investment decisions. While most of the technical indicators chosen are frequently used, and the goal is to portray a set of charts and outlook consistent with Wall Street, the indicators are selected based on our ability to interpret and explain them to investment professionals, policymakers from other fields, and the general public. Therefore, the viewpoint portrayed and any predictions presented as to future market performance imitate, but never replicate, those of any one private or public financial institution. They do not reflect the views of the Federal Reserve System, and they are published to improve public knowledge but are not validated with regard to accuracy of data or analysis and cannot be used for professional purposes." Yet you continue to assert that the report legitimizes technical analysis in some way. --Benna 00:56, 6 April 2007 (UTC)

First, I did not "allege" that the Fed's publication "legitimizes technical analysis." Second, the disclaimer in that publication does not change what I actually did say, namely that the Fed uses and publishes technical analysis. I also said these facts belong in this article about technical analysis. Third, I put back the mention of the Fed as the source of the study regarding fx support/resistance levels. The source should be named, whether it's the Fed, Harvard, or Podunk U. Rgfolsom 17:57, 9 April 2007 (UTC)

Further to this discussion, Alan Greenspan has stated in the past that economists cannot predict. It is not original "research" that the Fed (and other Central Banks) use technical analyis. It is a known fact that they use it, read it, and make decisions partially based on it. Ask any sell-side FX or fixed income technical analyst. Here is another Greenspan quote that says economists cannot predict, but where he essentially states the credo of every technical analyst (October 14, 1999). In it he states "the market price patterns remain the same". Although he implies that you cannot predict the bursting of the bubble, stating that the pattern is the same means either you can predict the bursting, or if not that, you can forecast the reversal. This is not research. This is exactly what he said. I was not saying that he uses TA (although I believe he does), but he is stating exactly the reason that technicians believe TA works:

"As I have indicated on previous occasions, history tells us that sharp reversals in confidence occur abruptly, most often with little advance notice. These reversals can be self-reinforcing processes that can compress sizable adjustments into a very short period. Panic reactions in the market are characterized by dramatic shifts in behavior that are intended to minimize short-term losses. Claims on far-distant future values are discounted to insignificance. What is so intriguing, as I noted earlier, is that this type of behavior has characterized human interaction with little appreciable change over the generations. Whether Dutch tulip bulbs or Russian equities, the market price patterns remain much the same.

We can readily describe this process, but, to date, economists have been unable to anticipate sharp reversals in confidence. Collapsing confidence is generally described as a bursting bubble, an event incontrovertibly evident only in retrospect. To anticipate a bubble about to burst requires the forecast of a plunge in the prices of assets previously set by the judgments of millions of investors, many of whom are highly knowledgeable about the prospects for the specific investments that make up our broad price indexes of stocks and other assets." Sposer 23:58, 5 April 2007 (UTC)

Stilted language and skepticsim

Saying "technicians 'try to' identify patterns" is to use a weasel phrase that implies skepticism of the method. You could say "try to" about anything a person does that could succeed or fail -- meteorologists "try to" forecast the weather, etc. -- but that's stilted prose.

As for "many" academic studies, this also implicitly sides with the skeptics -- unless the second half of the sentence also says "many" other studies say it may produce…, at which point it reads like a debate. The "some" vs. "other" construction is more neutral and natural. Rgfolsom 20:07, 11 May 2007 (UTC)

Saying "technicians identify patterns" is certainly not a weasel phrase - it's a statement of fact. Except that the 'fact' in this case is just plain wrong. Give me a weasel rather than a marketeer or a sockpuppet any day (and no, that's not intended personally). Pleclech 20:33, 11 May 2007 (UTC)

Extensive edits

I've edited for readability, and to remove redundant language. Also used quotes from sources with appropriate formatting. Rgfolsom 19:27, 18 May 2007 (UTC)