Author: Robert Hyatt
Date: 08:49:52 09/06/02
Go up one level in this thread
On September 06, 2002 at 11:29:01, Sune Fischer wrote: >On September 05, 2002 at 18:35:02, Uri Blass wrote: >> >>In that case you could use the 2 decimel numbers to claculate better estimate >>for the times relative to the published time. >> >>I wonder what is the reason that it was not done. >> >>Uri > >The number of decimals also indicate how sure you are of a number, you don't >just measure a number, and then write it with all the decimals you measured, >that would in fact be wrong. >You have to know how many of the decimals are accurate and how many are 'noise', >if you only trust the first decimal, then rounding to that decimal is the right >thing to do. > >When you read 2.0 you should know the 'real' number lies somewhere in the range >2.0 +- 0.05, it is implied. > >If you want more decimals you need to run more experiments, this could be a >problem if for some reason they didn't have enough time on the machine, or the >review of the article was done at a later stage etc. > >-S. Or, as I showed Martin, if the results vary so much from run to run that even that .1 digit is not very accurate. :)
This page took 0.01 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.