Systems and bear mkts

[Dennis Holverstott <dennis@xxxxxxxxxx>]

Tim ponders:
> I wonder just how well current systems will
> cope with vicious, prolonged downtrends? And further, if they don't cope well,
> how many systems developers will have the market experience/knowlege to develop
> new systems that cope well with a spiralling downtrend in the stock markets?

I think the answer is not many. Heck, most of the *commercial* systems
don't do very well a couple of months after they are written.

There are a couple of things system developers can do, at least for the
stock indexes. 

The index futures only go back to 1982 but there is daily data for the
cash Dow back to the early part of the century. So testing on that data
is a good exercise.

I agree with Tim and Neal that bear markets behave differently. Bull
markets seem to have slow up moves with sharp down spikes. Prolonged
bear markets seem to have slow down moves with sharp up spikes. One
thing you can do is "invert" your test data (x - price). See if your
system will trade a down trend as well as it trades an up trend.

-- 
   Dennis