You can try something like White's Reality Check or Hansen's SPA to see if
any of your 50 systems are really 'better' than any of the others.
Modifying these tests to use something else other than returns as the
objective to be maximized might be tricky, but doable.
If each of these 50 systems is calibrated every week or 2 as you indicated,
I would also recommend studying the PDC (parameter domain correlation) to
find out if the calibration is anything more than just picking the maximum
out of white noise. If the calibrations don't change much over time, this
is not a likely concern.
The documentation for the package "ttrTests" should help point you towards
references for the above tests. However, because of the way the package is
written, the code might not be particularly useful. I have been meaning to
update the package to make it easier to apply in situations like yours and
for a wider variety of data types. That project will move closer to the top
of my priority list if I manage to get a postdoctoral fellowship this fall
=)
Cheers,
-David
[[alternative HTML version deleted]]
_______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. If you want to post, subscribe first.
-- Also note that this is not the r-help list where general R questions should
go.
