I will discuss in more detail what I saw on Thursday, but thought a short discussion of my methods might be useful before I do that. My model is based on one relationship, and seems embarrassingly simple. Each wave breaks down into 5 smaller waves, and those smaller waves, when taken as points, exceed a certain correlation threshold, which is an R^2 value above .99. In the history of my posts, I have noted many, many waves, and each wave adheres to this rule. So the model is not complicated. The complication comes in correctly discerning the waves. So since stumbling across that relationship, I have tried glean what I could from others about correctly identifying waves.
Some waves are fairly straight forward, and are easy enough to follow. But waves can be infinitely complicated, as each can further break down into additional 5 wave sequences. The complicated the wave, the more difficult it can be to follow. it. As an example, I will look at the long term SPX chart.
Looking at the chart, Wave I began at 666.79, and ended at 1219.80. Taken as a point, it would be (666.79, 1219.80). Wave II would be (1810.10, 2872.87), and Wave III (2346.58, 3393.52). The R^2 value of the three points is .9925. So that would meet the requirements for completing a 5 wave sequence. And I guess the obvious question is how many false signals given? The answer is very few. Once the first and third waves are identified, a range for completing wave 5 can be given. If it turns out to not be wave 5, the wave will terminate either above or below that range. That is my model in a nutshell.
When I give targets, or what I have been calling primary targets, they are based on the range where wave 5 would complete to satisfy my model. At times I will mention secondary targets. While the above rule is absolute, and is true for every wave, I have always tried to find additional relationships. The holy grail of my model would be able to project each wave based on prior waves. So I have notice some other relationships that may usually work, but are not universal. So I try to distinguish between the two by using "primary" and "secondary" target labels. An example of this would be my 2030 secondary target for the completion of the wave down from 3393.52.
What I noticed was that the corrective wave generally completes based on previous points. If you take (Start of wave, End of correction), (End of correction, Wave III), and (Wave III, Wave V), again the correlation comes in above .99. So for this example, (666.79, Corrective Wave), (Corrective Wave, 2872.87), (2872.87, 3393.52). Now you can compute where the correlation is 1.00, and a likely target. For this example, the target is 2030. Sometimes the range can be large, but this method usually comes close. Should that target be broken, the next target would be computed by using the next lower high of the previous wave, or 2116.48.
All of what I do here is based on correlation of points. Some are experimental in a sense, so I want to make a clear distinction between those and hard Wave 5 targets. But I decided to discuss some of this to show that there is some method to the seeming madness, and not just numbers I pick out of thin air.
Which brings us back to Thursday. Given the previous wave set up, I think I mentioned Wednesday's 2280.52 low as a possible Wave 3. The implication being there wold be a move higher, followed by one more move lower to complete a 5 wave sequence from 3393.52. On Thursday we did get a move higher, with he SPX reaching 2466.97.
I consider waves to be fluid, reflecting the sum knowledge of investors, or speculators. So my labels are based on the most probable outcome given the wave structure. My view is that a change in structure reflects a change in sum knowledge. As long as the points themselves are not changed, adjustments to the count can occur up until a 5 wave sequence completes. So when I looked Thursday's action, I saw this:
Although the SPX moved higher through the day, it never seemed to have a lot of conviction. If the 2280.52 low was Wave 3, Wave 4 higher should have completed in a 5 wave sequence to the upside. When I looked at it more closely, I saw the first three moves higher as 2280.52 - 2359.75 - 2305.91 - 2409.81 - 2319.78 - 2431.03. (2280.52, 2359.75), (2305.91, 2409.81), (2319.78, 2431.03) has an R^2 value of .9962, and indicates a 5 wave sequence ended at 2431.03. The move from 2431.03 to 2385.83, when looked at on a smaller time scale, also breaks down into a 5 wave sequence. Then the move from 2385.83 to 2466.97 breaks down as (2385.83, 2432.27), (2408.31, 2456.50), (2421.08, 2466.97), which has an R^2 value of .9956, completing another 5 wave sequence. So the entire mve from 2280.52 to 2466.97 broke down into 3 waves, and not the expected 5. This implies that this move was part of a corrective wave that would mean a move lower below 2280.52. Then the corrective wave would complete with a move higher. Since we know the first and third waves, once we see the low (Wave d), we will be able to calculate the Wave e end point. That would only be the second wave 2553.93, meaning 3 more waves down to complete that sequence. And since that would be the third wave down from 3393.52, it implies still another wave down after that.
So with the wave structure as it is indicates some work to the downside still needs to be done before any rally would occur. But I'll take a closer look at that once we see how the current sequence plays out.
And most importantly, everyone stay safe and well.
Nice work.
ReplyDeleteWhat about C1-> C5
2.478,86; +/- 0,762; C1
2.711,36; +/- 0,618; C2
2.280,52; +/- 1,000; C3
2.466,97; +/- 0,50; C4
No going to
2.062; 1,00; C5
I like the fibs. Just have never been very proficient with them. But a lot of my targets do seem to match up with fib targets. Definitely something I'd like to look into.
DeleteHello Steven.
ReplyDeleteI am intrigued by your method. Just for fun I plotted the points (from 666 on) which you described above and let Excel derive the R^2 value. you stated.
I'm just curious, do you have some "feel" for why using 2 market price points, at different times and intervals of time, establish 1 point? Not criticizing, I'm only interested. Any new, innovative market analysis ideas interest me.
I'll keep an eye on your site if you choose to return comment. I'm not sure if the url I've provided as part of my profile here will be something you can respond to directly.
Thank you.