[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: TradeStation Precision - Summary



PureBytes Links

Trading Reference Links

On Thursday Bob Fulks wrote (in part):
 
> Since I recently spent a lot of time looking at this topic, I decided
> to summarize my findings so that I could find them again. I thought
> this might be useful to others so am posting it.
> 
> Bob Fulks
> 
> -----
> 
> INTRODUCTION
> 
> This message will summarize my experiments with TradeStation to
> determine the precision of arithmetic operations, illustrates what
> errors it may cause, and how you might "work-around" these errors.
> Please notify me if you spot any mistakes in this as I would like it
> to be as accurate as possible.
> 
> 

Today Bob Fulks wrote (in part):

> Interesting, but very fuzzy logic (pun untended, considering the 
> source). I have no more time to waste arguing with you on this.
> 
> YOU ARE WRONG! 


In between Pierre Orphelin wrote (in part with typo's corrected):

> Most of you forget what is the precision carried by the original data
> (price) and what is the precison expected from any calculation that
> derives from the original prices.


I stayed out of this because like Bob, I have no time to waste
arguing with PO and a few of his occasional cohorts about this
or any other topic.  We have been through this precision topic
many times in the past.  

To answer Bob's first question, I don't see any errors in what
he wrote on Thursday.  I agree that PO is wrong and/or misleading.

The part of PO's message which is quoted above is correct.
One should not expect more accuracy in a computed answer than
exists in the data provided to define the problem.  But you do
have a right to expect as much, or to be able to define how bad
it can be (worst case).

With 32 bit floating point precision calculations, it is VERY easy
to get totally meaningless answers.  I've spent most of my career
solving fluid flow problems.  I assure you, engineers don't have
the instrumentation to provide problem definition to 7 decimal 
digits of accuracy, but you would not dream of trying to model
the behavior with single precision computations.

The speed and memory "problems" are an exaggerated hoax created 
by PO and friends, as explained in the past.  I'm not doing it
again.  Hang in there, Bob.

Rod