__/ [Roy Culley] on Thursday 09 February 2006 02:29 \__
> begin risky.vbs
> Black Dragon <bd@xxxxxxxxxxxxxx> writes:
>> flatfish wrote:
>>> On Mon, 06 Feb 2006 03:57:58 +0000, cola-stats wrote:
>>> Finally some useful, accurate and unbiased statistics instead of
>>> that tripe that Roy Culley tries to pass off as "statistics" on a
>>> weekly basis.
>> Culley's stats amount to nothing more or less than disingenuous
>> trolling. I don't read COLA regularly, but I did notice some
>> comments about his stats being biased / manually edited.
> I clearly state what may not be included in my stats and why. Before
> I did that I posted an article on COLA asking what others thought.
> There was not a single objection if I remember correctly.
> IMHO my COLA stats simply exclude noise. Most newsgroups suffer from a
> few who wish to disrupt the group simply for the sake of it.
The extent of COLA, in terms of the number of posts (currently 11,485 in
my newsreader) per user/time unit, makes Culley's statistics helpful. Why?
* Let us begin with the fact that relatively few *distinct* people partic-
ipate in COLA. This observation and very reality makes it easier to reduce
the load to a manageable one. In KNode, I can easily tell when to hit 'i'
(ignore) on a new thread (probably 80% of all threads), so I need never
get involved in uninteresting discussions, let alone have them in sight.
Filtering based on sender is prejudice, which is why I glance before hit-
ting ignore. If in doubt, I can look up the stats or get an impressions of
them on occasions.
No cop is every perceived as a good cop unless he never uses the club and
handcuffs. Culley is in the misfortunate position of having to label some
people. Some labels, such as Bailo's, I do not agree with, which is why
his stats should be taken with a barrel of salt.
* Some people come to this Linux-oriented group without any purpose but to
offend and upset its regulars. If a fortnightly message helps us spot
trolls early on, why not use it? Trolls come and go and when I first
joined COLA I was replying to the wrong people. This potentially sent me
to people's killfile without realising it. I know the importance of ex-
cluding people like tab from discussions. *ANY* discussion. It is the only
way to make such 'noise' disappear naturally.
* more reasons omitted for brevity
> As for my stats being biased, given the reasons I explain what I
> exclude, why are they biased? As for 'manually edited', what do you
> mean? The fact that my script points out trolls? I don't manually edit
> other than add / remove 'comments' regarding trolls.
>> Since I've been posting stats for years in another group I figured
>> why not post some decent stats here too since it appears some people
>> would be interested in them.
> And what is the huge difference in your stats as opposed to mine?
> Other than the added reports I provide that turqstat doesn't there
> ain't that much difference.
ngstats.pl says much more. The only intersting bit in turqstat is the number
of posts from each individual. Readers will often overlook the rest because
they only aim higher where things are objectively measureable.
VISTA - Venereally-Infectious, Sexually-Transmitted Aliment
http://Schestowitz.com | SuSE Linux | PGP-Key: 0x74572E8E
5:00am up 23 days 0:16, 13 users, load average: 0.34, 0.49, 0.39
http://iuron.com - help build a non-profit search engine