Formal Studies in Technical Analysis

trader111 · #1 6th July 2008, 10:34 PM

Link:
http://papers.ssrn.com/sol3/papers.c...ract_id=603481

Scroll down to page no.72

Contains a set of about 10 tables which summarize research done till 2004.

trader111 · #2 6th July 2008, 10:52 PM

Quote:

There is no evidence at all that the analysts could generally distinguish random
series from the true price series. Additionally, the analysts were in general hugely
overconfident of their skill at doing so. While the submissions apparently did
well in terms of picking extensions that ended closer to the true extension than
should be expected, this is explained by a tendency to pick extensions that end
inside the range of endings. Technical analysts might hold out hope for the
effect seen with extreme answers.

Link to the paper:
http://www.burns-stat.com/pages/Work..._challenge.pdf

karthikmarar · #3 7th July 2008, 12:03 AM

oh... another TA basher....

trader111 · #4 7th July 2008, 12:13 AM

not just another TA basher

but yes, a critic trader.

btw the tables gives many studies which support TA

AW10 · #5 7th July 2008, 05:52 AM

Interesting research. There are 2 sides of any coin and it depends which side one is looking at.

oxusmorouz · #6 7th July 2008, 08:22 AM

Interesting but true. A lot of these "analysts" can't differentiate between random data and real stock data.

When I showed them this chart: (price is multiplied by 1000)

8/10 people started analyzing it by giving supports, resistances, fibo retracements, even trendlines and EW counts. Worse, one even inferred that it is a textile stock and started talking about how bad fundas look for textile cos. It in fact is purely random data.

Stationary throughout, lacks fat tails and is bluntly simulated (not even as complex as the one mentioned in the paper). Unlike the real one:

sudoku1 · #7 7th July 2008, 02:14 PM

Quote:

Originally Posted by oxusmorouz

oh god.....what a thrashing......this cant b ours....
looks like a hang seng or neikkei........

darsh_goswami · #8 7th July 2008, 04:29 PM

I was one of them Oxy ,.. LOL..

But for any TA Exists on this earth, he want ask what is the name of security, he will look at the price and will try to find out where its heading to.

I still remember how u sent me the chart and u asked whats support and resistance for this? if u question something ,u will get the answer for the same. Now who knows what that chart is all about? It was just a price movement for any TA. I wish ur getting my pt.

Regards,..
Darsh

trader111 · #9 7th July 2008, 05:39 PM

Quote:

Originally Posted by darsh_goswami

I still remember how u sent me the chart and u asked whats support and resistance for this? if u question something ,u will get the answer for the same. Now who knows what that chart is all about? It was just a price movement for any TA. I wish ur getting my pt.

ROFL
He sent the same chart to me too. How can you find support/ressitance when there are none?

Go well thru this one:
http://www.traderji.com/words-wisdom...illusions.html

trader111 · #10 10th July 2008, 01:17 AM

The article was written about 7 years back, probably in light of dimnishing returns in trend following strategies (simplest eg. Moving Average Crossover), in DOW. Neverthless it gives good insights about how MA based indicators may be rendered useless.

Quote:

HAS TREND-FOLLOWING CHANGED?

We have translated the general question into five approachable topics of interest:
♦ have general trend-following methods changed?
♦ have the sources of return for trend-following changed?
♦ have the magnitude or pattern of returns from trend-following changed?
♦ have the characteristics of trends changed?
♦ what are the implications for trend-following return characteristics?

1. Have General Trend-Following Methods Changed?
No, there is no evidence as such.

2. Have the Sources of Return for Trend-Following Changed?
No, there is no evidence or logical way to support this.

3. Have the Magnitude or Pattern of Returns from Trend-Following Changed?
No, there is no evidence supporting this during the past 10 years.... One caveat to the response to this question is related to the time horizon over which it is asked. If this same question were posed over a very long time frame, the answer might be yes, there is evidence. Performance in the past 10 to 15 years is lower than the results from the late 1970s through the mid 1980s.

4. Have the Characteristics of Trends Changed?
Yes, maybe. But is it significant? In recent years, the overall magnitudes of trends have been smaller. This reduces the theoretical edge of a trend-following model significantly.
....This is especially disheartening given that the vastly lower net return is accompanied by identical risks to the initial equity and to the open trade equity give back from peak valuation. This nonlinear efficiency is analogous to any industry with fixed and variable operating costs. Margins can expand significantly if the magnitude of a trend matches well the time frame of the operating model.
....The old stock market adage that markets fall twice as fast as they rise is an example of this behavior. By some measures, the speed of these retracements has increased on average during the last few years, sometimes reaching four to five times the rate of price movement with the trend.

5. What Are the Implications for Trend-Following Return Characteristics?
Many. While conditions described above persist (trends of smaller magnitude, faster countertrend price movements, difficult market groups) risk-adjusted returns for trend-following will be lower than they would otherwise.

More to the Point: Will Trend-Followers Change?

Events and experience stimulate evolution. Positive reinforcement induces behavioral changes.
...And finally, perhaps, after years of neglect, the powerful investor advantages that can be achieved through proper structuring will be harnessed.
...Yes, because of the successful maturation of mixed investment styles adopted by some already, in progress at others (such as our own firm), and inevitably to pressure others to move to diverse trend-following variations.

Link:
http://www.aima.org/uploads/2001%5CJun%5CWelton.pdf