|
PureBytes Links
Trading Reference Links
|
Sometimes data drop themselves and if one is not astute enough to pick
that up, the apparent results can be very misleading. When applying
statistical tests to market data or even backtest results one should
always question indications of statistical significance, especially if
sample sizes are large. In addition to potential misinterpretation
when tests are applied correctly (and almost every basic stats text
contains cautions about misunderstanding results from tests on large
sample sizes), there are also the limitations of the statistical tests
themselves, particularly if assumptions are violated. For example,
many tests assume observations are independent, identically
distributed (iid) and if that is not the case, hypothesis tests can
significantly overstate degrees of freedom. In biostatistics for
instance, effective sample sizes for huge "randomly" drawn samples are
sometimes tiny due to local homogeneity. Market data and output
derived from it can easily violate the iid assumption due to stock
sectors, nonstationary serial correlations, etc. So I'd say
interpretation is more than occasionally necessary, it's essential.
--- In amibroker@xxxxxxxxxxxxxxx, Franco Gornati <francogornati@xxxx>
wrote:
> quanttrader714 wrote:
> > Franco, what you said about variance is absolutely correct but
first,
> > I don't believe there is a "true" value.
>
> and I too ..
>
> > And second, larger is not always better because large sample sizes
> > can make small differences that are not important statistically
> > significant. So interpretation can be critical.
>
> yes, but if you drop data you lose information. They may not add a
lot
> but reducing them surely doesn't improve your position.
> But if you add non pertinent or non omogenous data then you may
need
> some interpretation.
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/GHeqlB/TM
---------------------------------------------------------------------~->
Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
|