|
PureBytes Links
Trading Reference Links
|
Clyde, I believe you have approached the problem in a logical manner and
have constructed a reasonable set of statistics to back up your hypothesis.
FWIW, your study agrees with something which I remember Ed Kasanjian posting
to RT a couple of years ago.
That said, I find it difficult to reconcile the distribution of your
statistics with what I observe in real-time trading of the markets. Time and
time again I see price pause and/or reverse at key fib retracement and
expansion numbers. There are just too many times where I have exited within
a couple of ticks of a 162% fib expansion target only to see price reverse,
or entered a tick above a 38% retracement only to see price breakout, not to
believe that these numbers are real. A couple of years ago I posted a series
of intraday trades in real-time all of which were fib based and most hit the
money. Perhaps a year ago, I posted a series on Symmetry again using fibs
and many of those worked out well. Note that I am not claiming anything
approaching perfection with fibs ... I have enough losing trades to prove
that also.
I am wondering if there are multiple sets of behaviors in the market which
depends on time frame and the more recent availability of computer based fib
calculation tools? Thus I tend to think in terms of another universe which
co-exists with the universe you studied. I wonder if you could run a similar
statistical distribution on intraday data (5, 15, or 60 minute) for the past
several years in a widely traded futures index such as SP or ES? Ditto for
just 3-5 years of most recent daily data.
There is another factor at work here which defies simple testing ... the fib
trades I do are selective based on the overall feel and pattern of the
market. Thus, the discretionary trader will use fibs for entry/exit but only
when there is a discernable and high probability pattern and trend showing
in the market. This is something which is second nature to the experienced
discretionary trade and I can not imagine how this can be simulated by
unfiltered swing statistics.
Earl
----- Original Message -----
From: "Clyde Lee" <clydelee@xxxxxxxxxx>
To: <realtraders@xxxxxxxxxxxxxxx>
Sent: Saturday, April 13, 2002 7:39 PM
Subject: Re: Re[6]: [RT] Fibo predictions
Let us get back to the original thesis of this discussion:
Are FIB levels any better than RANDOM levels.
The analysis I have provided indicate that NEITHER holds
an advantage over the other.
It is easy to pick one or two examples which seem to indicate
that there is significance in FIB levels but if we look at some
1500 to 2000 pivots over a period of 72 years or so there is
clear proof that FIB levels are no more significant than some
kind of random selection of numbers.
This analysis has been run from very minor swings to very
major swings and the same distribution exists.
I know that there are people who swear by FIB as a guide to
pivots and have used their concepts successfully but the fact
is that there is no more significance to FIB levels than some
random set of numbers we select.
Clyde
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
Clyde Lee Chairman/CEO (Home of SwingMachine)
SYTECH Corporation email: clydelee@xxxxxxxxxxxx
7910 Westglen, Suite 105 Office: (713) 783-9540
Houston, TX 77063 Fax: (713) 783-1092
Details at: www.theswingmachine.com
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
----- Original Message -----
From: wavemechanic
To: realtraders@xxxxxxxxxxxxxxx
Sent: Saturday, April 13, 2002 7:03 PM
Subject: Re: Re[6]: [RT] Fibo predictions
Clyde:
I think you have oversimplified a fairly complex problem. The goal is to
identify the major turns within perhaps +/- 1-2%, as in the case of similar
studies (e.g., pitchforks, So9, etc.). In order to do this, traders usually
use more than just fibo levels to make a judgement as to what the
significant level is. They look for meshing between major fibo levels and
other subjective/objective factors (e.g., EW, indicators, etc.). Take a
look at the attached. Would your analysis pick up the major turns on this
MSFT daily, rejecting the minor swings within each leg? That is the primary
goal. My guess is that you will have to use different pivot parameters to
catch major and minor turns. I also suspect that consistent results will be
very difficult to achieve using pivots alone. I do know, however, that more
often than not the major turns are in synch with the "standard" fib, So9,
etc. levels. In addition, there appears to be reasonable evidence that the
patterns are fractal, extending up/down in time. Of course, to see that one
has to change the "microscope's magnification" by, for example, adjusting
pivot "+/-" size, indicator settings, etc., as a one size fits all would
most probably not be suitable in the majority of cases.
Bill
----- Original Message -----
From: Clyde Lee
To: realtraders@xxxxxxxxxxxxxxx
Sent: Saturday, April 13, 2002 4:13 PM
Subject: Re: Re[6]: [RT] Fibo predictions
What does it take to prove that Fib numbers have no validity in
estimating where prices might make a turn.
I have posted several charts which clearly prove this is the
case and yet we continue to hear BS about Fib numbers.
Just to make it clear, here is an explanation of the method
behind the attached chart.
Consider a swing such as the following"
b
/ \
/ \
/ \
a \
\
c
and calculate the ratio (b-a)/(b-c)
or a swing of the following type
a
\
\
\ c
\ /
\ /
b
and calculate the ratio (a-b)/(c-b)
Do this with a mathematically definable method
of picking swings and accumulate the ratios in
a spread sheet, sort the data by ratio, and make
a chart.
The attached is exactly that chart for the S&P
index from 1930 until now using an 8 bar length
window for picking pivots.
A careful examination will indicate a more or less
CONTINUIOUS distribution of ratios of the 1713 swings
which existed in the period of study and had a ratio
of less than 2.0.
If there were ANY VALIDITY to the concept of turning
of prices at FIB levels then there would be a bunching
of data about the various FIB levels and not the very
continuous distribution that is found in the data.
Again, people may use the fib levels as levels at
which to be aware of potential turns in direction of
prices but the analysis says that we are just as well
off with a random set of lines since there will not
be any grouping around them either.
Please, examine these data in detail and if I am
missing something then provide the data or interpretation
of these data that says otherwise.
Clyde
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
Clyde Lee Chairman/CEO (Home of SwingMachine)
SYTECH Corporation email: clydelee@xxxxxxxxxxxx
7910 Westglen, Suite 105 Office: (713) 783-9540
Houston, TX 77063 Fax: (713) 783-1092
Details at: www.theswingmachine.com
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
----- Original Message -----
From: ztrader
To: realtraders@xxxxxxxxxxxxxxx
Sent: Saturday, April 13, 2002 1:37 PM
Subject: Re[6]: [RT] Fibo predictions
On Saturday, April 13, 2002, 11:10:12 AM, wavemechanic wrote:
w> Are you thinking that confluence is associated with Fibs only?
w> Could it be extended to any coincidence of indicators?
w> As long as the indicators are truely different (e.g., velocity vs
acceleration, etc.).
How about a 50 ma and a 200 ma with identical values, and price is
approaching this value? Would this 'confluence' have more importance
than if the two ma's had quite different values?
ztrader
To unsubscribe from this group, send an email to:
realtraders-unsubscribe@xxxxxxxxxxxxxxx
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
To unsubscribe from this group, send an email to:
realtraders-unsubscribe@xxxxxxxxxxxxxxx
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
Yahoo! Groups Sponsor
ADVERTISEMENT
To unsubscribe from this group, send an email to:
realtraders-unsubscribe@xxxxxxxxxxxxxxx
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Stock for $4
and no minimums.
FREE Money 2002.
http://us.click.yahoo.com/orkH0C/n97DAA/ySSFAA/zMEolB/TM
---------------------------------------------------------------------~->
To unsubscribe from this group, send an email to:
realtraders-unsubscribe@xxxxxxxxxxxxxxx
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
|