Author: Robert Hyatt
Date: 19:51:59 12/10/02
Go up one level in this thread
On December 10, 2002 at 22:09:48, Jeremiah Penery wrote: >On December 10, 2002 at 21:17:04, Robert Hyatt wrote: > >>On December 10, 2002 at 20:37:37, Jeremiah Penery wrote: >> >>>On December 10, 2002 at 20:15:15, Robert Hyatt wrote: >>> >>>>On December 10, 2002 at 13:25:55, Jeremiah Penery wrote: >>>> >>>>>I'm not trying to prove anything. But for you to say Vincent is hand-waving is >>>>>a bit much, since he's actually referring to some real data. This is 6 >>>>>positions worth of data, when all the tests you did the other day were only >>>>>using 5 positions. So you claim one test is meaningless, and the other isn't? >>>> >>>> >>>>Several positions, run _several_ times, and averaged together to smooth out >>>>the variation... >>>> >>>>Big difference... >>> >>>http://www.talkchess.com/forums/1/message.html?269805 >>> >>>You ran 4 positions. Nowhere did you say that you ran them multiple times or >>>that the NPS figures are averaged at all. >> >>I _always_ run them multiple times. I have stated this dozens of times here. >>And I have given the reasons why I do this... _many_ times... > >The text in your message implies that you ran them once, and there's absolutely >nothing there to indicate multiple runs. Just because you usually, even always, >do something, doesn't mean it should automatically be assumed that you did, when >the implication is the opposite. > There is _no_ "implication" in my post of any kind that I can see. If you look at my discussions about parallel search and testing for the N years I have been posting here, I have _always_ mentioned the necessity of running a test multiple times, and I have even posted raw data showing how much variability there is in a parallel search using the same position, and just running it over and over. I'd _hope_ that it is not necessary to continually re-state the obvious, over and over, every time I post _additional_ parallel search data. I _always_ run tests multiple times and average the results. I have a simple shell script that does this automatically so that I just type the command and check back in a few minutes to see the average of the results... >Even still, the fact that you ran only 4 positions, and yet complain about the >data Vincent is using, that uses 6 positions, is pretty silly, no matter how >many times you ran your positions. When vincent doesn't even _run_ the tests? and I am "pretty silly"? Just call me "silly boy" then... if that is your opinion... Several post one set of results. One person posts another. And _that_ post is the "bible"??? Not in my world...
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.