Re: Fibonacci numbers in the gold market (help please)
- From: "Nick" <tulse04-news1@xxxxxxxxxxx>
- Date: Mon, 14 May 2007 15:56:11 +0100
<iamorasa@xxxxxxxxx> wrote in message
news:1178995397.582118.169340@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Hi,
I'm a newbie & a non-mathematician.
I trade gold futures markets, and have come across an interesting
mathematical relationship between the 'ranges' - that is, the runs up
from a low point to a high point (also known as a bottom and a top).
Gold has been rising in price since the early 2000s. Like all markets,
it has done so in stages. I.e. it will come off a low, climb up to a
high, then drop back to another low. (But the recent low is not
usually lower than the previous one, which of course is why the
general trend is up.) These low-to-high ranges are usually very
visually obvious on a chart. You ignore smaller ranges within them:
it's the ranges lasting some months that are of interest.
If you look at all the up-ranges going back to mid-2003, the largest
by far was the 'parabolic' final one from the base in March 2006 to
the May 12 2006 top. (The highest high in 25 years, which has not yet
been exceeded.)
That range (SpotV NYMEX futures chart, for any traders reading this -
it closely parallels the actual gold price) was $197.50.
I decided to multiply 197.50 by each of the Fibonacci numbers over 1,
sequentially, and see what I got. I started with 1.236, then went to:
1.382
1.5
1.618
1.786
2.0
Each calculation of course gave me a number. I then deducted that
number from the high of May 2006 - which was $732. (The 26-year high
where my final, big range ended.)
I then scanned back through the chart for three years. In order, the
numbers resulting from the above calculations were very close to a
significant low. The average deviation was $3.73. Given that the lows
varied widely in price - from $340 to $534 - that was fairly accurate.
There were only 7 significant lows in that three-year period -
unambiguously defined low points on the chart, from which runs up
began. The above six Fibonacci numbers hit close to all 7 of them.
(One low occurred at the same level twice (a 'double bottom').)
So the first 6 Fibonacci numbers over 1.00, applied to the final range
in a series of ranges, 'predicted' (backwards) the starting point of
every other range preceding it - in order. (Reverse order to be
precise.) It didn't miss one of them, nor did it predict any which did
not materialise.
Any general comments would be welcome.
However my specific question is: How could I have picked this pattern
up before it ended - preferably near the start of the series of
ranges?
As it stands, my final range of 197.50 'predicts' with reasonable
accuracy the lows which preceded it. Is there some corollary to that
formula which would allow me to do the reverse?
I've avoided using actual data here to keep it simple - but am happy
to send/post it on request.
NB: I don't seem to get any results from using Fibs of the first low
to low range and the first low to high range. I.e. by using these
early (2003-4) ranges as my 'primary' range instead of the final
range. Their predictive power is limited.
What I need is a way to reverse engineer the final range, because it
works. I.e. pick up the pattern earlier.
By the way, I would have thought that this is in the area of forecasting
which is more the subject of sci.math.stat.
Nick
.
- References:
- Fibonacci numbers in the gold market (help please)
- From: iamorasa
- Fibonacci numbers in the gold market (help please)
- Prev by Date: Re: Paths
- Next by Date: Re: Three parallels do not cut - one should assume, shouldn't one?
- Previous by thread: Re: Fibonacci numbers in the gold market (help please)
- Next by thread: Condition distribution
- Index(es):
Relevant Pages
|
