Recent content by ivanbaj

The need for prediction is reversely proportional to the amount of information one has. Imagine 4 individuals in the middle of a dark night. 1. Old one. That one observed sun rises for 4 billion years 2. New born. Has not seen a sunrise yet. 3. 1 day old. Seen one sunrise 4. An alien on...

<img src="http://www.elitetrader.com/vb/attachment.php?s=&postid=3247424">

<img src="http://www.elitetrader.com/vb/attachment.php?s=&postid=3247422">

<img src="http://www.elitetrader.com/vb/attachment.php?s=&postid=3247419">

<img src="http://www.elitetrader.com/vb/attachment.php?s=&postid=3247417">

<img src="http://www.elitetrader.com/vb/attachment.php?s=&postid=3247411">

This league is going places. Any one interested to invest? http://www.youtube.com/watch?v=gZcSEE7TuGU

What is a trend line? I think it is in a form of p(t) = a + b(t) That is not going to work.

More prove:

And here is why:

This is a chart of the close price for EURUSD 10/17/1994 till 7/3/2001 The Deutsche Mark is used as a proxy for the years before 1999 How much would you believe this trend to make a prediction for 7/4/2001????? If you offset the line up you can make it a trend line or a channel if...

I believe that I can prove that trend lines and channels are useless. But I don't want to stick my neck out here. It will get cut in a mil. second.

The problem with TA is that Y(t) = a + bt Is not appropriate for most financial time series. x(t) and x(t+1) are quite correlated. A simple acf() from R will show you that. QQ plot and a histogram on the residuals from the lime model will show a substantial trend pattern.

These work: <img src="http://upload.wikimedia.org/wikipedia/commons/3/3a/Astrologyproject.svg" /> <img src="http://upload.wikimedia.org/wikipedia/commons/7/79/12_houses_of_heaven.jpg" /> <img src="http://upload.wikimedia.org/wikipedia/commons/d/d1/Zodiac_woodcut.png" />