I was (unsuccessfully) trying to find results on what the distribution of the MACD values for a stationary time series with IID returns would be. Are there any such results or any that go in a similar direction?
$\begingroup$
$\endgroup$
3
-
$\begingroup$ Hi: I can't remember if it's been done for MACD but the procedure generally involves converting the original rule into a rule about returns. The paper at link is kind of relevant ( like I said. I can't recall if they consider MACD ) but there are also others also that could be more relevant. papers.ssrn.com/sol3/papers.cfm?abstract_id=2604942 $\endgroup$– mark leedsApr 16, 2019 at 17:49
-
$\begingroup$ you may find my answer here relevant: quant.stackexchange.com/a/44914/29443, it basically doing what @markleeds suggests but in a different context $\endgroup$– Attack68 ♦Apr 16, 2019 at 19:18
-
1$\begingroup$ @attack68: Hi: what you did at that link is interesting and related. My take is that an estimation-minimization approach will depend largely on the time-frame and how often you re-estimate. if you re-estimate often and, if the time frame is too fine ( even days ), I have a feeling your estimates of the weights will be very unstable. but maybe not :). neat idea for sure. the approach in the paper gets a closed form for MA but I can't remember if it does it for other ones. $\endgroup$– mark leedsApr 16, 2019 at 23:55
Add a comment
|