|
PureBytes Links
Trading Reference Links
|
Mats,
I just returned from traveling to catch your e-mails on TradeStation's
precision. I couldn't agree more. Try using the coeffR function to
calculate a correlation. Correlation, StDev, Regression all require
squaring large numbers. This is a disaster with TS's single precision.
With Currentbar > 1,000 the difference between TS's calculation of a 90 bar
correlation versus Excel or C++ (which agree to 9 decimals) is .32 versus
.56. I call that significant. To add insult, Omega adds code to constrain
the calculation to +/- 1. Correlation is by definition constrained tp +/-
1. The only reason to constrain it is to hide the fact that you can't
calculate it correctly.
On an even simpler scale, calculate a simple 1000 bar moving average in TS
and compare to Excel or C++. TS can't even do this with precision to ONE
decimal place.
Neal
-----Original Message-----
From: Bengtsson, Mats [mailto:mats.bengtsson@xxxxxxxx]
Sent: Wednesday, July 18, 2001 10:51 AM
To: pierre.orphelin@xxxxxxxxxxxxxx; omega-list@xxxxxxxxxx
Subject: RE: I think I have to change my opinion from bad precision to buggy
c alculations
Now you are kidding again, good humor. Paying you for EL consulting must be
considered a bad way to waste money. If I had 1000000000000000000000002$ and
gave you 1$ would have to be considered a bug :-)
I notice you are once again off topic, so you are out of arguments. And that
is natural, it is hard to keep claiming Tradestation has 7 digits precision,
which is better than single precision, and at the same time try to explain
why calculations that come out correct in single precision does not even
come close in Tradestation. One has to be quite stubborn to even try... :-)
Or spending a great amout of arguments trying to explain why 10.3-10.2 comes
out with a rounding error in Tradestation but not in languages (except maybe
fuzzy logic), that utilises the math processor and IEEE storage format. And
still claiming that this is very good, is better than the languages that
deliver the correct results, and that this has something to do with physics
lessons taught at some kind of school. I do not recommend you to read a
whole book, that would be asking too much. Just sit down and think (avoid
fuzzy logic): If one program (Tradestation) arrives at different results
than for example C++ If those results arrived at by this program
(Tradestation) are incorrect due to lack of precision. If the actions
performed by this program (Tradestation) are illogical due to the lack of
precision. If the write protected code delivered by Omega does nothing to
circumvent these numerical errors in indicators? Is it really a good idea to
keep claiming Tradestation has higher precision than what is claimed by C++?
Is it really a good idea to defend Omegas indicator code as logical
according to physics lessons? Is it really a good idea to keep claiming that
this is what should be expected from an expensive calculation program?
Just think, not too hard, just a little. :-)
|