The forensics algorithm for the 10s



#2

I had a good weekend for collecting by finding a Pickett N-500S in very good condition for two dollars, a old paper reactance slide rule by Shure for a dollar, and a HP-10s for three dollars.

I ran Mike Sebastian's forensic algorithm on the 10s. The answer is 9.000000002124 . The eight zeroes after the decimal point makes the result better than any other HP machine listed at Mike's site. Am I doing something incorrectly or is the 10s really that good?


#3

Quote:
The answer is 9.000000002124 . The eight zeroes after the decimal point makes the result better than any other HP machine listed at Mike's site.

The HP-30S, also listed at the site, returns 9.000000000, for which Rodger Rosenbaum has provided an explanation:

http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/archv015.cgi?read=85973

Edited: 19 Jan 2010, 10:48 p.m.


#4

The end of Rodger's dissertation states:

Quote:
So finally, this is why the HP30S gets exactly 9.00000000000000000 when running the calculator forensics test. To see the real error, one should subtract the input argument of the test from the final result. That is, do:


arcsin(arcos(arctan(tan(cos(sin(9))))))-9


On the HP30S, the result is 0, leading one to believe that the HP30S is *really* accurate. But redo the test as:


arcsin(arcos(arctan(tan(cos(sin(9.1))))))-9.1


to foil its rounding near integers and see a result: 4.833288903E-11


Not all that great for a calculator that can do 24 digit calculations, eh?


I did the Rodger's 9.1 version of the forensics algorithm on my 10s and received 9.100000002602 . Does that suggest that the 10s isn't doing any of the funny arithmetic that Rodger was writing about?

#5

Quote:
Does that suggest that the 10s isn't doing any of the funny arithmetic that Rodger was writing about?

Yes, apparently the 10s isn't doing any trick to make it look more accurate than it really is, otherwise it would return 9.00000000000 in the original test. Assuming the algorithms are correct, it is possible internal calculations are carried out with 15 or 16 digits internally. On classic 12-digit HP calculators intermediate results are rounded to the number of digits of the display, so the forensic results will be a bit away from 9, even though the algorithms are flawless:

http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/archv016.cgi?read=103151

It looks like HP is back to its original philosopy (see Katie's result for the 30b below).

#6

The 30b comes in at 8.99999864267, the same as the Pioneers (20S, 32S, etc.).

-Katie


Possibly Related Threads...
Thread Author Replies Views Last Post
  HP 10s+ Easter Egg (Statistics) Eddie W. Shore 1 215 04-11-2013, 04:25 PM
Last Post: Andrew Feist
  HP 10s+ hardware Mic 1 196 04-07-2013, 07:52 PM
Last Post: Eddie W. Shore
  Review of the HP-10S+ Mic 0 149 12-30-2012, 09:06 AM
Last Post: Mic
  HP 10S+ order on DeRekenwinkel.nl Mic 2 234 12-06-2012, 04:44 AM
Last Post: Klaas Kuperus
  Question about a "combinations" algorithm Namir 9 537 09-20-2012, 04:51 PM
Last Post: Gilles Carpentier
  Linear Programming - Simplex Algorithm LarryLion 5 335 09-04-2012, 10:57 PM
Last Post: David Hayden
  HP 10S mixed up keys Iqbal 0 142 05-19-2012, 08:54 AM
Last Post: Iqbal
  Algorithm for fitting a logistic curve? Tim Wessman 5 365 11-13-2011, 01:22 AM
Last Post: Wes Loewer
  HP-15C LE calculator forensics? Joel Setton (France) 19 986 09-11-2011, 08:01 AM
Last Post: Dieter
  A fast and compact algorithm for the normal quantile Dieter 13 758 04-22-2011, 07:11 PM
Last Post: Paul Dale

Forum Jump: