Fourier long term

Been working on the fourier, worried about the long cycles. Using 92 week fundamental cycle. Where the daily chart posted was looking for a 5/6 high and 5/13 low, weekly cycles are also looking for a low in three weeks. No good feeling about the size of this plunge, it doesnt look to bad on the daily but looks bigger on the weekly. Weekly analysis agrees with daily surge into June but weekly projects longer to the end of June. Next week is going to be a good top, watch out for 5/13 week. Will short the indexes for that one. From May 13 week up strongly to end of June. Flat and choppy for four months, and up hugely from Oct until December. That's pretty easy, out next week. Long from May 13ish through June, long again from October. Will post both short and long charts next week as top approaches.

Ok, here it is.

This is based on three 92 week cycles, cycles are changing so less accurate
over a 176 week average.

Why is the current rally stronger than projected, cycles are changing.
The magnitudes of the top 3 cycles are dropping off precipitously,
only the 13 and 23 week cycles that will produce the June rally are growing.

Thus the bear mkt is over and less volatility forward.

Therefore the May 13 lows will not be as bad as projected here, the
daily chart posted before with moderate low is more accurate.
It should be a W4 type action.

And here is the cycle magnitude spectrum, cycles dying off. Was an echo of the 2003 low.

It will take some development to have a new set of long term cycles.

Oh, it's SPX log.

Ok, here it is.

This is based on three 92 week cycles, cycles are changing so less accurate
over a 176 week average.

Why is the current rally stronger than projected, cycles are changing.
The magnitudes of the top 3 cycles are dropping off precipitously,
only the 13 and 23 week cycles that will produce the June rally are growing.

Thus the bear mkt is over and less volatility forward.

Therefore the May 13 lows will not be as bad as projected here, the
daily chart posted before with moderate low is more accurate.
It should be a W4 type action.

What date was the last low at 1651.3?

My methods are looking for a low May 6th and a rally into May 12, quite the opposite to yours.


blu

I certainly can be wrong. It's just mathematic best fit, the possible errors can come from wrong fundamental cycle selection, or cycles changing in real time. I use the phase spectrum to identify the presently stable fundamental cycle. That may be key to my system working thus far. If the calculated fit to actual price is reasonable, it's the best I can do by this method. I posted the chart of the magnitude spectrum, there is also a phase spectrum. I want the phase angles to be stable through the last fundamental cycle. It's fairly good when the 45 and 135 days turned up and called those lows in real time, I posted here. I'll have to look up that data again for you blu, I dont have that file with me now. Check back later.

That was 8/7/06

The weekly analysis is based on the last three 92 week fundamental periods.
94 week may also be valid.

I had posted a daily based fourier looking for a 5/16? high but it didn't feel right
and I kept working it. I changed it last weekend with some posts on the
subject.

The daily is using 3.5556 144 day fundamental periods with quite a good fit as
it nailed this weeks ramp. This changed my analysis from a 5/16 high to 5/13 low.

I use fourier in physical engineering and process application. There's a lot to it
and I could go quite a bit deeper. I think what was wrong with the earlier daily
analysis was the presence of imaginary cycles. These are cycles that are there
only for the purpose of adjusting other cycles that dont fit just right.

In physical application it is non linear. I've worked with weighting techniques
based on correlation coefficient.

The key to the weekly low dying off is the decay in cycle magnitudes, sort of
an echo of the 2003 lows, a remnant of that cycle dying off. Something new
is there in the longer cycles that can't be measured yet until there is one full
fundamental cycle ( another 2 year?). So I'm weighting the daily periods for now.

What date was the last low at 1651.3?

My methods are looking for a low May 6th and a rally into May 12, quite the opposite to yours.


blu