Author: Ralf Elvsén
Date: 12:33:31 09/07/02
Go up one level in this thread
On September 07, 2002 at 10:27:27, Robert Hyatt wrote: >On September 07, 2002 at 09:25:57, Ralf Elvsén wrote: > >>Keep as many digits as your machine allows you until the bitter end. >>Then you make an estimation of the uncertainty, and present a properly >>rounded value together with the uncertainty. Rounding before this >>stage is a no-no (although I'm not sure that was what you intended >>to do). >> >>Ralf > > >It depends on circumstances. IE in my dissertation, I had the freedom to >run each test hundreds or thousands of times, which I did. For the DTS paper >that was not an option. With a single position, precision to 2 decimel places >is pointless as the test logs on my ftp machine shows... I don't want to start a long argument, but one has to be careful. If you have (say) a long series of numbers which vary wildly and you want to present them as an intermediary step in a calculation, of course one want to see 1.41 rather than 1.414214... But the number that goes into e.g. calculating an average should not be rounded. First of all as a principle. Use what you measured, not something else. Secondly, doing the rounding and justifying it is a a needless complication. To keep all digits in memory or in a file is trivial. Then you have the cases which might not be so clear cut. It's easier to stick to the most precise way, which also happens to be the easiest. I thought you subscribed to the KISS-philosophy? :) Ralf
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.