Forum Archive Index - September 2003

["david.gibson" <david.gibson@k.co.nz>]

>That's not an assumption. If you read Malkiel's "Random Walk Down Wall
>St", you'll see that he and his students tested TA empirically - and
>found that it performed no better than buy and hold, or after brokerage,
>even worse. You can argue that their choice of signal might have been
>wrong, or that the state of the art has improved since the 80s, but
>their criticism was based in actual evidence, not just reasoning from
>first principles.

Just to be clear ... Technical Analysis is the process of spotting trends
and other patterns in a timeseries and using this in a predictive capacity.

The time series analysis techniques I employed, in ancient times, are more
aligned with econometric modelling.  We did find some variables that did
provide signficant correlation with future market price trends.  Mainly
these were to do with interest rates.  However, there were some other
independent variables that performed as market bell weathers.  In the end -
the project was successful, we did build a useful predictive model.

However, I am very aware of the limitations of the tools of the time.  (More
appropriately, the assumptions upon which the underlying statistical
techniques were based).

Today, there are some very promising emerging techniques based on non-linear
dynamics (catastrophe theory, chaos theory are derivatives).  Again, I would
not call these techniques TA - they are really extensions of econometric
modelling.

Best Rgds
/dbg


----------------------------------------------------------------------------
To remove yourself from this list, please use the form at
http://www.sharechat.co.nz/chat/forum/