Author: blass uri
Date: 22:22:16 07/04/00
Go up one level in this thread
On July 04, 2000 at 18:16:34, Robert Hyatt wrote: >On July 04, 2000 at 05:51:27, blass uri wrote: > >>On July 03, 2000 at 21:37:29, Bruce Moreland wrote: >> >>>On July 03, 2000 at 07:00:45, Chris Whittington wrote: >>> >>>>On July 02, 2000 at 12:15:05, Bruce Moreland wrote: >>>> >>>>>On July 02, 2000 at 07:23:04, Chris Whittington wrote: >>>>> >>>>>>On July 02, 2000 at 05:21:45, Bruce Moreland wrote: >>>>>>> >>>>>>>Nobody is going to sit down and do psycho-statistical analysis of a thread's hit >>>>>>>count drop-off rate. >>>>>> >>>>>>Tres drole. Sorry about the lack of accents. >>>>>> >>>>>>Most persons, with half a brain even, would be able to instantly intuit sensible >>>>>>conclusions from readcount data. Quite why you would imagine it necessary to >>>>>>perform SSDF-like statistical analysis to several places of decimals presumably >>>>>>with degrees of confidance figures attached escapes me. Or are you arguing on a >>>>>>reducto absurbam basis? I thought I already made it clear the idea of having the >>>>>>data was just an aid to avoid the danger of being too arbitrary and subjective. >>>>>> >>>>> >>>>>I think the data would be useless. The interpretation is everything, and it's >>>>>likely that the data would support any interpretation. >>>> >>>>There is no such thing as 'useless' data. As you point out, it is all in the >>>>interpretation. I think what you actually mean is that the interpretation would >>>>be 'useless'. >>>> >>>>Actually, you don't even mean that, since 'useless' is an inappropriate >>>>descriptor. You mean that the data would be interpreted _unscientifically_ and >>>>that such an interpretation would be not 'useless' but 'unhelpful'. To whom? >>>> >>>>Presumably you would not attempt to argue that a _scientific_ interpretation of >>>>the data would be 'useless'? Probably 'contentious' from your viewpoint, but not >>>>'useless'. >>>> >>>>> >>>>>I don't see how you can object to the possibility of using math on numbers, but >>>>>I doubt there is any math that would help, either. >>>> >>>>I don't object. I was mildly counter-mocking you. >>>> >>>>> >>>>>>>If the data were recorded with intent to use it as an aid >>>>>>>in helping out with moderation decisions, all that would happen is that 1) The >>>>>>>moderators would use the data (any data) to bolster their existing opinions >>>>>>>about topicality, and 2) People who aren't moderators would use it in order to >>>>>>>bolster complaints: "You said my C++ post is off-topic but the page was >>>>>>>refreshed 13 times." >>>>>> >>>>>>Your point (1) shows a degree of cynicism to the moderators and their decision >>>>>>making processes. >>>>>> >>>>>>Your point (2) shows a degree of cynicism to members and their tendency to >>>>>>complain. Or their motivations for doing so. >>>>>> >>>>>>I share your frustrations. >>>>>> >>>>>>> >>>>>>>Sorry, but I don't agree that the intended purpose of this feature is necessary >>>>>>>or desirable. >>>> >>>>Who-whom? >>>> >>>>I expect you to challenge my motivation for the idea. Obviously you are >>>>suspicious. But such thoughts can be turned to face at yourself, I think. >>>> >>>>To repeat: >>>> >>>>"The >>>>>>>moderators would use the data (any data) to bolster their existing opinions >>>>>>>about topicality, and 2) People who aren't moderators would use it in order to >>>>>>>bolster complaints:" >>>> >>>>Have you become so cynical that group (1), the moderators would >>>>_unscientifically_ abuse the data; and that group (2) the not-moderators would >>>>do the same? >>>> >>>>Group (1) plus group (2) equals everybody, doesn't it? >>>> >>>>Are there any 'good guys' here, in your view? >>> >>>Imagine you had suggested that we get a rock, and send it to Steve. Every time >>>there is a new post, Steve is to lick the rock. If the rock tastes salty, Steve >>>tells the moderators to delete the thread as off-topic. >>> >>>I would argue that there is any logical correlaction between the taste of the >>>rock and the topicality of the post, so that any data produced by Steve is >>>useless. >>> >>>I think your idea is similar to the rock-licking idea, only worse, because >>>nobody would believe that rock-licking is a good idea, but Uri Blass thinks that >>>hit counting is a good idea. >>> >>>bruce >> >>I believe that there is a correlation between the topicality of the post and the >>hit count. >> >>I also do not say to use only hit count to decide and in more than 90% of the >>cases I am sure without using hit count if the post is on topic or off topic. >> >>Uri > > > >if you use this logic, you will conclude that automobile accidents are great >things to happen. Just watch what happens in the south-bound lanes on a major >interstate when there is a wreck in the north-bound lanes. Does the 'hit count' >there mean a wreck is good, or bad? Bad things attract people just like good >things. As bruce said, a hit counter is interesting, but a high one doesn't >mean a thread should be kept any more than a low one means a thread should be >deleted. I think that 'hit count' is independent of thread significance... I did not think to use hit count to decide if to delete insulting posts but to decide if a post is on topic or off topic in cases that it is not clear(for example book on C++). In most of the cases it is clear for everyone if the subject is on topic or off topic and I think that in cases that are not clear the question should be how many people here read it because the idea of not posting off topic posts is that we do not want to see a lot of posts that nobody is interested on. If something is only about chess than it is clear that it is off topic because there is another forum to post it. If something is only about computers but may be relevant to chess it is not clear because there is not a third forum for these posts(there is a forum for c++ but if the asker wanted to know about good book to help you to write a chess program then this is a different question because graphic for example is not relevant to building a good chess engine). Uri
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.