|
PureBytes Links
Trading Reference Links
|
While the typical moving average, RSI, or stochastic are extremly unlikely
to run afoul of floating point problems, statistical functions seem to be
particularly vulnerable to single float problems and divide by zero errors.
(On more than one occasion I've dropped the standard Bollinger Band
indicator onto a chart and gotten a divide by zero error.) Also, converting
prices to log values provide the ability to optimize trading systems for net
profit without being skewed by a few large and/or recent trades - single
float calculations performed on log prices are particularly prone to float
errors and other anomalies.
Earl
-----Original Message-----
From: VISTA Research and Trading <skt@xxxxxxxxx>
To: James F. Mazzulla <tagteam@xxxxxxx>; omega-list@xxxxxxxxxx
<omega-list@xxxxxxxxxx>
Cc: Steven McQuaide Tennis <mcquaide@xxxxxxxx>
Date: Wednesday, September 23, 1998 7:51 PM
Subject: Re: TL_Easy Language Mathematical Precision
>
>James,
>
> Floating point values (real numbers) are indeed stored as four byte
>(single precision) reals.
>
> I once approached Perry Kaufman with the question posed below. His
>response was that there are virtually no calculations we perform on price
>data that require greater precision.
>
> In the occasion that greater precision is required it is possible to
>perform them in an external DLL using double precision values but, indeed,
>the return value must once again be a single precision number.
>
>Samuel
>
>
>At 04:43 PM 09/23/98 -0400, you wrote:
>>From: tagteam@xxxxxxx (James F. Mazzulla)
>>To: sci@xxxxxxxxxx
>>Cc: omega-list@xxxxxxxxxx
>>Subject: Re: TL_Easy Language Mathematical Precision
>>Date: Wed, 23 Sep 1998 17:10:28 GMT
>>
>>On Wed, 23 Sep 98, Bob Brickey wrote:
>>>Even though it is possible to print more precise-looking numbers on the
>screen, TS4 and TS5 both perform mathematical calculations using what is
>called "Single Precision" in Visual Basic. Easy Language function
>arguments are passed as Single Precision values. Results are returned as
>Single Precision values.
>>>
>>Thanks for the clarification, Bob. Can you tell us if this is
>=particularly= problematic and, if so, what, if any, course of action needs
>be taken? TIA
>>
>>Best regards, Jim
>>
>
>
>
> Samuel K. Tennis VISTA Research and Trading, Inc.
> 8103 Camino Real C-409 voice: 1(305) 273-1321
> S. Miami, FL 33143 fax: 1(305) 273-9164
> skt@xxxxxxxxx http://www.vista.com/
> ***** Easy Language Spoken Here *****
>
>
|