|
PureBytes Links
Trading Reference Links
|
Here is my first attempt to predict the next
bar Low.
The statistical results are not bad at all.
In the attached gif, next bar candles are plotted together
with todays Support prediction for tomorrow
[the last candle is missing, since it is not known
yet.]
The support line is green when we predict tomorrows Low with
less than 2% error
For ^NDX it was possible 148 from the last 200 bars [74%] ,
from the end of last Aug.
For AMGN it was the 84%
71 stocks of the N100 database obtained less than 2% Low
approximation for more than the half of the 200-bar test period.
How about a less than 1% error between the predicted and the
actual Low?
For ^NDX it was possible the 48% of the last 200
bars.
9 stocks were between 50% and 60% of this period.
The decay rate was not dramatic, 34 stocks were "less than 1%
correct" more than the 40% of their trading bars.
Instead of various artificial [and many times unrealistic]
Support levels [previous supports, Fibonacci supports, even trendlines
supports]
is it better to *know* next bar Low with less than 2% error
for the majority [71%] of the stocks ?
Think about...
Dimitris Tsokakis
Yahoo! Groups Sponsor
Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
Attachment:
error2.gif
Attachment:
Description: "Description: GIF image"
|