Recently, after much effort on the part of his critics, marketsurfer finally gave up his definition of trend:
"A series of moves in one direction that can be quantified and exploited as deviating from randomness."
He had previously declared stock trends to be nonexistent. I'll leave it to you to see the connection.
My own recent epiphany is this:
How you define a trend depends a great deal on the context in which your trend occurs.
In the context of financial data, I go with the classic trader's definition:
A trend is EITHER a series of higher highs and higher lows OR a series of lower highs and lower lows.
In the context of coin-flips, I define a trend as EITHER 3 or more heads in a row OR 3 or more tails in a row. Yes, surf, trends can be and in fact often are generated by purely random processes. So your "deviation from randomness" is a nonstarter.
In the context of criminal investigation, a trend is two sufficiently similar crime scenes ("You're a detective now, son. You're not allowed to believe in coincidence anymore." -- Commissioner Gordon)
In all contexts, the definition of trend should be all about "pattern-ness". A trend is first, foremost and always a pattern. Bringing in extraneous factors like exploitableness just muddies the water.
First establish the pattern. Only then should the focus be shifted to how best to exploit it.
"A series of moves in one direction that can be quantified and exploited as deviating from randomness."
He had previously declared stock trends to be nonexistent. I'll leave it to you to see the connection.
My own recent epiphany is this:
How you define a trend depends a great deal on the context in which your trend occurs.
In the context of financial data, I go with the classic trader's definition:
A trend is EITHER a series of higher highs and higher lows OR a series of lower highs and lower lows.
In the context of coin-flips, I define a trend as EITHER 3 or more heads in a row OR 3 or more tails in a row. Yes, surf, trends can be and in fact often are generated by purely random processes. So your "deviation from randomness" is a nonstarter.
In the context of criminal investigation, a trend is two sufficiently similar crime scenes ("You're a detective now, son. You're not allowed to believe in coincidence anymore." -- Commissioner Gordon)
In all contexts, the definition of trend should be all about "pattern-ness". A trend is first, foremost and always a pattern. Bringing in extraneous factors like exploitableness just muddies the water.
First establish the pattern. Only then should the focus be shifted to how best to exploit it.