|
PureBytes Links
Trading Reference Links
|
This is a good example.
If a simple moving average [of prices] gives different results in MS TS and
Excel, you cannot conclude that Excel was right because of its better
precision.
Knowing that MS, TS are single precision, that a single precison is far
enough to respect the 5-6 digit precision in prices, MS, TS and XL should
give the minimumm accuracy ( eg the averages are correct at the requested
precision that prices dictates).
The only explanation is that the garbage part beyond 6 digit is not the
same in all of these programs, and that some of the decisions were taken
from these irrelevant decimals.
Now, I could agree that the garbage part in XL is of best quality than the
TS garbage part beyond the 6th digit, but garbage is still garbage, eent if
it is the best garbage in the world.
And if the results of such a system are so sensitive to such small
meaningless difference, this also means that the system is garbage itself.
Could you expect that a such unstable system will produce results showing
any form of generalisation on price data, where the intrinsic noise is far
above the minimum precision of prices???
For my 15 years of physics teaching, I remember that it took years to some
of my students to learn to stop to write the meaningless extra digits given
by their programmable palm calculators.
After all, I still have some years to change your mind, but unfortunately,
I cannot rate ( note, evaluate) your copy, so it will be more difficult for
me.
PO
> -----Message d'origine-----
> De : MikeSuesserott [mailto:MikeSuesserott@xxxxxxxxxxx]
> Envoye : dimanche 29 juillet 2001 00:52
> A : Bob Fulks; omega-list@xxxxxxxxxx
> Objet : AW: TradeStation Precision - Summary
>
>
> "We had such a case some time ago on the Metastock list where one
> poster got
> into big arguments with Metastock support because the software would give
> him signals different from those he was getting with Excel. The reason
> turned out to be some decimal in a calculation which finally lead to a
> different result, caused by different floating-point
> implementations [single
> vs. double precision]. And these calculations were just simple Moving
> Averages, not Fourier analysis or the inverting of Hilbert matrices!
>
> "So this issue of accuracy is not to be taken lightly, IMHO."
>
> Best wishes,
>
> Michael Suesserott
>
|