[eastbay]

[ QUOTE ]
[ QUOTE ]
What is "the trend" and how does one "follow" it? I'm not looking for hand-waving here, I'm looking for testable, quantitative, repeatable measures for finding "the trend" and testable, quantitative, repeatable methods for "following" it.

Any of you trend followers want to start?


[/ QUOTE ]

Some hands involved, but its mostly an "eye" thing! [img]/images/graemlins/smile.gif[/img]


[/ QUOTE ]

Ok, now that we've gotten the hand-waving out of the way, let's get down to the real business of defining a trend.

What I am after is a quantitative, testable definition of a trend. The reason for this is that this is the only way to do a legitimate experiment to see if "following" (to be defined next) said trend is in fact profitable over the long term, in a verifiable way, not "so-and-so said he does it and can show you his audited statements" or "I do this every day and make a living at it" or other such assertions which are not independently verifiable.

The closest thing you provided to a workable definition is:

"If the price is moving up over time, then it is in an uptrend" and vice-versa.

But that's a pretty imprecise definition. How much time? How consistently up? Monotonically up, or just more up than down, or simply up overall?

So the first question is: how do we select "uptrend" data given a price stream in arbitrary time and price units?

Is 9,10,11 (say N-min bar closes) an uptrend? Why or why not?
Is 9,8,10,11 an uptrend? Why or why not?
How about 9,8,12,11,13?

What is the actual criteria that separates uptrend from non-uptrend?

eastbay