Yes Larry, I agree stops are important with any "system"....... My bollingers are very wide. 2.618 in order to reduce the chance of false signals. And this is ONLY for markets that are trading in a small range. It is for days you can pick up 10 cent moves when it is virtually trendless.
Size of BB will not alone reduce false signals enough. There has to be a filter to accomplish that. Walter
In my model time is activity derived from volume. So my model has time in axis. Before better read the guide http://perso.wanadoo.fr/harrytrader/_sgt/m1m2_1.htm http://perso.wanadoo.fr/harrytrader/dji/211102/cours_Plan_SPOT1h.swf For example, today theorically 8873 was the break zone. Standard error is 2 to 3 point and maximum standard error is 8-10 points and Future must also pass this level to validate the break. If it failed to break it should do it within two theorical time units. Statistically my time units is about twice the real time unit. So it should occur within the first hour.
I had an instructor in college that got me interested in trading and he used some strange stuff to predict turning points in time. I'm not saying any of it works, because I gave up on it years ago, but he seemed to do well with it. He would take two major highs (or lows) (could be in any time frame, and he had rules that I don't remember for knowing what counted as "major") He counted the number of bars between these two points and then multiplied it by 1.618 (Fib ratio) and added the result to the lowest low (highest high) between the turning points. This would give him a projected bar in the future where he expected a turning point. He said the more significant the highs (lows) used, the more likely that the projected turn would be accurate. He was also big into planetary stuff, and that ephimerus wheel in the link that nkhoi posted earlier was a mainstay in the classroom. He was also a big believer in Welles Wilder's "Delta Phenomena", a book I bought on his recommendation that details a methodology for projecting turning points using moon phases. Its a good read and comes across very convincingly until I tried to make it work in real-time. I now just consider it a $175 tuition payment in my trading education. I just couldn't get any of this stuff to work with enough confidence to convince me it was better than random. I tend agree with those that say you cannot predict the timing of turns with any real accuracy, but that is just my personal experience after a lot of research and "tuition".
To WarEagle, To take two highs and low between is a starting point of timing approach to the TA. If one is to apply constant ratio of 1.62 or any other constant, it will not work simple because distance of two major point is not only one to be taken into equation. Size and speed of moves between significant points has to be considered too. I do it for Bonds and for Eminis with 75% or better accuracy. Walter
Fib Ratios are used for things like that in a simliar way as Fib Ratios are used for price retracements, but not so many people do that, so perhaps that could give you some edge (?). A good book that elaborates on that is 'Dynamic Trader' by Robert Miner. I have it, but never used it for anything really practical, although I must say that he is very good at predicting turning points. He nailed the last SP500 (October 10th) bottom to the day, but so did my mechanical systems (all of them gave a long signal on that day, which does not happen that often), except that he did it a few days ahead of me.