Standard deviation is the measure of scatter of a group of data. A standard deviation of 1 or 2 datum is meaningless. I don't think that there is any adaptive standard deviation per se.
If you look at a Bollinger Band the location of the upper and lower bands are based on the standard deviation of the close for a given number of intervals. You could develop a standard deviation where at each close the number of datum used to calculate the standard deviation could depend on another factor like volatility or volume or whatever you want to use. You can download a FREE book on statistics. Look up SticiGui. If you can't find it, you can view a series of lectures on statistics at http://webcast.berkeley.edu/course_details_new.php?seriesid=2009-D-87441 <http://webcast.berkeley.edu/course_details_new.php?seriesid=2009-D-87441&se mesterid=2009-D> &semesterid=2009-D Lionel Issen From: [email protected] [mailto:[email protected]] On Behalf Of Rob Sent: Wednesday, June 16, 2010 6:37 PM To: [email protected] Subject: [amibroker] Adaptive Standard Deviation Hi All, I'd like to code an adaptive standard deviation... I think I use the right term. Lets pretend todays first interval is bar one, I'd like to calculate a standard deviation for each bar based on the number of bars we've had so far today. The StdDev() function clearly only takes a fixed period for it's calculation. In this example the 'period' would increase by one each time we get a new interval. Any ideas...? Am I going to have to use a loop and calculate it manually...? Thanks
<<image003.jpg>>
<<image004.jpg>>
