1st of all, I'm as mere as mortals come. I'm just using what comes easiest to me as I try to map out the heavens where the real immortals reside. The perplexity of my work to you and others is the very same thing I feel about the work of many others.
Here's an edited response I once sent to someone who inquired about my methods. Maybe it'll help you see where I'm coming from...
Thanks for your compliments and interest. "Par-Trend" is an oscillating design engineering product, period. My thing works with price and time, but makes no attempt whatsoever to define time-specific cyclic behaviour. It's just another tool for quantifying trend. I've never found one that plotted price objectives (I call them demands upon price) for a given trend scenario, so I'm simply taking a stab at it myself. It's all experimental, under development and subject to changes until further notice. I never see it evolving beyond what it is, which is a tool to systematically or methodically plot price points for a full range of trend scenarios. From those scenarios, one is left to investigate further and ultimately judge for himself the path of least resistance, using other TA tools. At least, that's where I'm at with it so far. A mystical discipline for forseeing the future? Not a chance, ever. A process of elimination for screening out lower probabilities? A definite maybe.
Most trend analysis focuses on peak and trough points. Kind of like starting from the outside and working your way to the core. Nothing wrong with such stuff. It's just that there's not much out there that focuses from the center and works its way out from there. Well yes, there are plenty of oscillators, but they remove price from the indicator by the way they're constructed, and an oscillator's intuitiveness seems to get lost with that design. Also, other oscillators don't really project forward by themselves. I wanted an incremented trend oscillator that visually followed price around on a chart (like a Bollinger band), and would project totally stable envelope lines into the future. That way, any given degree of trend you'd want to investigate would have its requirement upon price, for any given time, distinctly plotted out on a chart. There was a day when I realized that a simple moving average consisted of two dynamic parts and one static part: (1) a new price being added, (2) an old price being dropped, and (3) everything in between that was nothing more than the remaining exact fractal of the moving average's time period. Another way to look at it is that the two dynamic "book-end" prices simply kick around the intact fractal between them, day after day. One thing led to another... I realized that I could therefore plot prices for unbiased straight-line slopes using envelopes, and I could set up any kind of increments I wanted (I chose Fibonacci) within a trend range defined by maximum slope achievements. And that's "Par-Trend". The envelopes follow the prices required to maintain a totally flat and horizontal moving average slope. They follow along in "percentile parity" with the oscillator's "zero point", or core price, with defined boundaries and benchmarks for the trend (moving average slope), hence the abbreviated name "par-trend".
There's nothing to prevent one from adding "outside/in" overlays like trendline or EW work to help determine how feasible a geometric idea is within the trend's price structure. Anyway, it's really nothing more than a navigation chart that shows where prices lay for any given trend vector, where the trend vector is the slope of a simple moving average. So what differentiates "Par-Trend" from other TA methods? I guess it's this: I can define the reasonable outermost parameters for future price activity, calculate within that a spectrum of benchmarks with Fibonacci increments (or selective benchmarks from the historical records), all for further input into and comparative analysis with other indicators.
As far as targets are concerned, I typically perceive them as moving objects subject to change. Time is the dynamic that makes them change. Momentum is the function of price over time. So essentially, par-trend studies are momentum studies. (end of response)
And much of that focus is upon the potential for divergences that may not be seen until the actually occur. The early influence of those already planning for these divergences (accumulation and distribution) mostly goes unaccounted for, so I'm just trying to fill that void in my own way.
The charts I present in Stockcharts format plot out this Fibonacci grid of a moving average's slope range (the 20ma is the most essential, and the one I usually post). I usually don't show my other overlays (like trendlines, wave counts, cycle partitions, etc.) because the charts get way too busy-looking and overwhelm the bare essence of what I'm trying to present. Besides, others do that better than me anyways, or most others can see where they would rather place them.
Here's a Stockcharts 50ma par-trend chart for NDX. Many institutions and investors tend to lean on 50-day results, so this might be worth noticing. As far as the 50ma slope is concerned, we nearly did a full pendulum momentum swing from the March low. We are still above the 50ma's 76.4% slope prices.
Here's a chart from my Trademan program, where I can combine and average out the aggregate momentums of charts like this from 1 day to 50 days. The plotted prices are from the 20ma's 61.8% slope, same as those shown on my first charts on this thread (they are a good proxy for continuation of the NDX's recent trend). But I'm using a par-trend summation indicator for 1 thru 50 days instead of 1 thru 20.
Now I can guarantee you that many institutions use far more sophisticated means to look for upcoming divergences between momentum and prices in the future, especially those promising some tenacity. But I'll bet they do look at it and plan for it. I can't say how or when they (and others) will act upon it, but they do see it and sooner or later act upon it.
I'm going to start adding NDX to these posts.
