Fooled by Technical Analysis

Solas0077

Active member
Messages
236
Likes
14
This is a new book by Michael Harris? Contents look interesting. Has anyone read it already? Price is a bit high and one can only access files from the author 's website, which is a kind of non-standard mehtod.
 
This is a new book by Michael Harris? Contents look interesting. Has anyone read it already? Price is a bit high and one can only access files from the author 's website, which is a kind of non-standard mehtod.

I would suggest that you ask Sharky if T2W would add it to their list of books to review. If both he and the author agrees then there would probably be a number of free copies for review.
 
TA's not a magic nor an easy approach. If you apply it poorly, you'll get poor results. Same applies to writing books.
 
TA is in fact difficult. It is an interesting book that focuses on what can go wrong with TA. Overall a good read with many examples of systems and patterns. Note that Harris is at some point blasting the expectancy formula and this is an interesting juncture in this book.
 
Straw man.

It's not difficult to debunk something that exists primarily because of the way in which one defines it. "TA" is an excellent example, given that it can be defined in so many ways. And given that Harris charges $2000 for his software, one can't help but wonder if his take on all this is chiefly a means of offering to "save" the trader from the evils of whatever when one can save himself at little to no cost.
 
Straw man.

It's not difficult to debunk something that exists primarily because of the way in which one defines it. "TA" is an excellent example, given that it can be defined in so many ways. And given that Harris charges $2000 for his software, one can't help but wonder if his take on all this is chiefly a means of offering to "save" the trader from the evils of whatever when one can save himself at little to no cost.

Now it is for less :)

BTW you cannot blame a man for being honest. I have read the book and it made a big impression on me. Nothing like I have read in the past. I think it will save me a lot down the road.
 
Financial markets are pretty close to being random walks, so that making forecasts using technical analysis is very difficult and of limited benefit. For example, if a trading instrument is, say, 98% random, and 2% deterministic, how can one expect to make significant profits using technical analysis?

The best way, in my opinion, is to carefully assemble a portfolio of instruments in which the random component is minimized, and the deterministic component is maximized, thereby allowing better forecasting ability. This is the idea behind cointegration, which is the method I use for trading. I assemble (and continually update) a portfolio of cointegrated ETFs, and obtain buy & sell signals from a Bollinger-band-like algorithm (actually, a Kalman filter applied to the portfolio weightings). This is a generalization of pair-trading, applied to 3 or more trading instruments.
 
It's only random if one doesn't understand how an auction market works.

As I posted earlier:
"Analysts" have been making livings for decades (centuries?) because their audiences seek certainty. And their audiences are willing to pay extraordinary amounts of money for certainty. But there is no certainty, only probabilities. The best -- Wyckoff, Magee, Mamis, Douglas -- have understood this. But the audience is, for the most part, deaf.​
Understand how the market works and none of it is random. It is all intentional.
 
It's only random if one doesn't understand how an auction market works.

As I posted earlier:
"Analysts" have been making livings for decades (centuries?) because their audiences seek certainty. And their audiences are willing to pay extraordinary amounts of money for certainty. But there is no certainty, only probabilities. The best -- Wyckoff, Magee, Mamis, Douglas -- have understood this. But the audience is, for the most part, deaf.​
Understand how the market works and none of it is random. It is all intentional.

Markets are "random" in a mathematical sense, which can be demonstrated objectively. That doesn't mean you can't beat the market, but it takes more than a mechanical application of technical analysis. The best traders use all kinds of information to beat the markets.
 
Markets are "random" in a mathematical sense, which can be demonstrated objectively. That doesn't mean you can't beat the market, but it takes more than a mechanical application of technical analysis. The best traders use all kinds of information to beat the markets.

That one can't predict the outcome of any particular trade does not mean that one can't predict the outcome of a series of trades. Whether or not it "takes more" than a mechanical application of technical analysis is another matter which is not necessarily related to randomness. Success also takes more than a mechanical application of fundamental analysis.

Randomness and unpredictability are not the same thing.
 
Randomness and unpredictability are not the same thing.

To the degree that a financial time series is random, it is unpredictable. The very definition of randomness is "the lack of pattern or predictability in events."

If you want to "beat the market", i.e., beat a buy-and-hold strategy, you have to make forecasts about future prices. To the extent that a trading vehicle is random, to that same extent its future prices cannot predicted. Equities are not purely random, but they are mostly random. That's what limits the effectiveness of conventional technical analysis when applied mechanically.

There are many ways to trade profitably, and they all involve overcoming this randomness problem. Some people use fundamental analysis, behavioral models, algorithmic trading, or their own intuition. My approach is to try to forecast a carefully assembled (cointegrated) portfolio, in which the randomness is reduced, and the deterministic component is increased. I'm not trading a stock or bond or currency, etc., but rather a portfolio of trading instruments whose net value can be forecast with higher confidence than any of its components separately.
 
The "randomness problem" exists only in the minds of those who believe that the market is random.

But if you're successful with your approach, keep on keepin' on. :)
 
The "randomness problem" exists only in the minds of those who believe that the market is random.

But if you're successful with your approach, keep on keepin' on. :)

We use what we know. My background is physics/math/programming, so I'm partial to quantitative methods.
 
"fooled by ta" is vague to say the least. Ta is only one piece of a process that makes money. It serves to aid direction potential and help plan risk management. So to analyse it on its own would be a very stupid thing indeed
 
Markets are mostly random but if you know the probability distribution you can make money. If not, do not bother.
 
Markets are mostly random but if you know the probability distribution you can make money. If not, do not bother.

You need to know a lot more than the probability distribution. (It's approximately normal, except with "fat tails".) However, an understanding of statistics can be very helpful, especially with quantitative approaches.
 
You need to know a lot more than the probability distribution. (It's approximately normal, except with "fat tails".) However, an understanding of statistics can be very helpful, especially with quantitative approaches.

If you know that some distribution has positive skew than you can make money, maybe a lot of money. All you need is a simple strategy to take advantage of that.

BTW the Harris' book is very good. Opened my ueys.
 
I would never trade a method that isn't profitable on a randomly generated chart.
 
Top