|
PureBytes Links
Trading Reference Links
|
List,
I really don't have the foggiest idea what the
"Bollinger Band System" is (whether provided by Bollinger or the Equis
add-on). The following discussion is an overview of how I have manipulated
the formula during the last few years. Since there seems to be a growing
interest in this area, I thought I might "spill" a few ideas, on the forum, and
see if any list-members can improve and contribute ideas to this
framework.
First, let me say: among all of the public,
"talking head", commentators, I consider John to be one of the good guys.
After attending graduate biz school, at an advanced age, I was delighted that
there was a practical application involving one of the few things I learned
while attending school: standard deviations.
Of course, we all know, I think, that those crooked
little lines that tend to bracket the price range are points on the chart that
represent (in John's formula) two standard deviations from the mean. I
like that concept, but I hate those crooked lines. As some of you know,
I've taken the information contained in the Bollinger formula and changed it to
allow the information to look a bit more logical (my subjective
opinion).
Instead of the crooked lines (the daily adjustment
representing the two standard deviations from the mean), I prefer to convert the
crooked lines into straight lines and chart the standard deviations as parallel
lines. In this restructured formula, the price is plotted as a percentage
of the distance between the chosen standard deviations (see attachment
"1").
The "new, improved" formula is:
<FONT face=Arial
size=2>((C+2*Std(C,20)-Mov(C,20,S))/(4*(Std(C,20)))*100)
I suggest adding horizontal lines at 0 (zero) and
100 (and now, a trader can compare and see that we are just representing the
information in a different "light").
Next, it's an easy step to implement a "system" ...
for example:
{Enter Long positions when the indicator
closes below 0Enter short positions when the indicator
closes above 100}
Enter Long:
<FONT face=Arial
size=2>Cross(0,Ref(((C+2*Std(C,20)-Mov(C,20,S))/(4*(Std(C,20)))*100),-1))
Enter Short:
<FONT face=Arial
size=2>Cross(Ref(((C+2*Std(C,20)-Mov(C,20,S))/(4*(Std(C,20)))*100),-1),100)
This "crude" approach returns 130 cents and
has an 10 & zero track record when forced onto
the December Wheat '01 contract (see attachment "2").
I suggest optimizing variables within the
formula. It's easy to explore results by varying the "trigger" levels, the
mean, or the standard deviations.
If you want to apply the "approach" to day trading
the naz or s&p, try stepping down to the "ever-popular" ten
period-ema. Use the 10ema and experiment with a stepped down version that
explores penetrations of only one standard deviation.
The combinations are endless. This approach
is just that: an approach. Filters, trend identifiers, combinations and
modifications are all possibilities. This
is not the holy grail. It just is a jumping off point for those who would
like to explore the realm of price (buried in an average) and it's relationship
to standard deviations.
Hopefully, this will stimulate a trading discussion
that will benefit our forum. If we don't get back to the purpose of
this forum, it might as well fold up and blow away.
Take care,
Steve Karnish, CTACedar Creek Trading<A
href="http://www.cedarcreektrading.com";>http://www.cedarcreektrading.com
Attachment:
Description: "2.gif"
Attachment:
Description: "1.gif"
|