Unless you know a way to predict trends, you can't detect one until it exists. The ker depends solely on the length of your lookback period; you can make it as short as 3 if you can stand all the false alerts.
I'm not sure if you're saying no trend can be exploited in that second paragraph. Observation shows that trends are exploited all the time, buy-and-hold being a prime example for most years since the Depression. As far as 'trending now', use a filtered price velocity if other methods are too laggy.
trend=expectation of the underlying process putting a value on the trend through statistical determination requires an infinite sample, that's basic probs.
So according to your definition, there's no such thing as a trend, because there has never been an infinite sample of data collected even if we add up all of human history. That's ... unique.
Keep it simple. Use the ADX with DMI http://www2.barchart.com/support/learning.asp?code=BSTK&what=adxdir
Thanks for introducing a very important intermediate/expert topic. The combination of the comments here on failures to be able to handle this and the major failures inherent in DBP's commentary do scope and bound one of the most important building blocks for intermediate/expert trading success. After a person is nailing the bulk of the market proffered intraday opportunities, where are the places he goes to be prepared to then extract other opportunites? The general sets of considerations dwell on recognizing the moves in market operating points and then doing a comprehensive rapid characterization of the new operating point. Obviously, these two considerations are what fouls up most back testing in several ways: the first set is omitted almost always; the second set is continued to be used well after that operating point is no longer applicable; and the timing of when to be doing each of the sets is not within the grasp of the person primarily because he does not have in his belief system three factors; they are: Migration of distinct operating points; how all operating points are arrayed; the overlap of gross characteristics re operating points. Finally there is the determinant of change. All of this, from an intermediate and from and expert viewpoint hinges on a paticular set of market knowledge that did not come up in the discussion to date. You, from your vantage point (beginner), nee to skip backtesting and go to market concepts. Go to this conceptual area for now. Learn about the pace of markets completely. There are several integrated principles. Once you deal with each, you get to see and understand how their dynamic works individually and, then collectively (meaning relative importance as well) you are able to see two facets: the changing relative importance and the more or less universal arrival and departure of each element that contributes to pace. From the above, you can understand why a person can be stuck in a place where that place, for him, generates a particular type and quality of response. All leraning begins with the market and not with tools that process data. To learn about the market you study the market. as a human and to extend your capability, you select and use tools to extend your mind's reach. For fun and after work, you can play with tools. If by chance you wee formally introduced to particular tools, always keep in mind that you learned about them because it is what you like to do. What you want to do and what you want to play with. For making money, the market TELLS you what tool to pick up. Pick it up. Find out what other attachments to that tool are called for. There are no tools mentioned in this thread so far by anyone and, especially, No advise from a formally trained person who has tools he likes to play with. This post from me to you is a "keeper"; it applies to a lot of things. I will knock out an expert level 80 pager on this soon. It is a good idea to drop some stuff into ET once in a while to act as a wake up call for the troops. Good luck and thanks again for your thread topical post it has been a long time coming.