Apr

22

To the extent which moving averages may convey information, there might be a better way to model the decay of investor (emotional) memory decay. Presumably t mattered more than t-1, which mattered more than t-3.

Many biological processes scale logarithmically, so here is an lnMA of daily range which weights each day less by ln(countback) (here for n days). r = range = SPY (H-L)/C:

nD lnMA = {r[0]/ln(2)+r[-1]/ln(3)+….r[-n]/ln(n)}/n

The most recent day's range (r[0]) is scaled by ln(2) to avoid the divide by zero problem.

Attached is simple 10d MA for SPY daily range, vs 10d ln MA, vs 1/10000 SPY close (I used 1/10000 to put it on same scale).


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search