Ways of testing Mean Reversion

Quote from Daring:

Markets typically revist most moving averages, but they have a higher chance of doing so when the moving average, within itself was trendless.

Eventually, all moving averages will be re-visited...because they are 'averages'.

However...

If you are looking at a moving average on a 'Daily' basis, you have to realize it is calculated on the 'Close'. Here's the problem - it may not reach the moving average on an intra-day basis...but at the end of the day, it will certainly look as though it did. This is why looking at charts can be deceiving.
 
Quote from oldtime:

well, that's what I am saying. Let's say you have determined that the market chops 85% of the time and trends 15% of the time. It will never help you determine what it is going to do next, but once it violates it's historic time average, you can start betting on reversion to the mean.

and really, isn't that all trading is? betting on history

they don't write about people that do it year in and year out

the only ones you hear about are the ones that bet "this time is different" (or more correctly, the "one")

Yeah, back before 2008 I used to make nice money doing this with Lehman. It was a beautiful thing. I still mourn its untimely demise. For the methods I use, there was never a more perfect stock. Sigh.
 
Quote from TheMagican:

Where?Can you please show it?


Overall good discussion in the thread. If you would like me to point out particular posts that I really liked.

1) There was a post which had an attachment that compared residuals of spline against MA. I found it useful.

2) There were some posts by Mike805 wherein he distilled his experiences. I found those posts useful too.
 
You don't need "fancy" tools to make mean reversion work, as some of the posts/threads links suggests. What you do need, is a sideways market that extends itself to confluence.

Crazy how people complicate this business so much.

Sometimes the brilliance can simply be found in simplicity.
 
Back
Top