|
PureBytes Links
Trading Reference Links
|
> -----Message d'origine-----
> De : Bengtsson, Mats [mailto:mats.bengtsson@xxxxxxxx]
> Envoyé : lundi 30 juillet 2001 13:53
> À : MikeSuesserott@xxxxxxxxxxx; pierre.orphelin@xxxxxxxxxxxxxx;
> omega-list@xxxxxxxxxx
> Objet : RE: TradeStation Precision - Summary
>
>
> I would suggest trying to avoid sin/cosine or any other cyclic function.
> Best thing would be to use real data (both operating on the same data
> series, like djia for example) whicheverybody could get their hands on of
> they wanted to repeat the experiment, and then use something that
> is on the
> line of what everybody tries to do in one way or the other
> (signal to noice filtering in one way or the other).
>
The sine cosine function produces a synthetic variable price that is
correct up to the 7th digit , and allowed me to highlight the propagation
error effect.
Using price data to test the difference with TS implies that I must use an
other program to have a benchmark, what I have not.
PO
|