OK

In the dark days after the election of 2004 despair and despond ruled in sunlight and in shadow.  As I recall, we all sat in our comfy chairs, empty beer can in hand, the 5 year old television in front of us seemingly permanently darkened.  The holiday season came and went.

Then, a new year, 2005.  Miracle of miracles, rejuvenated and revitalized participation.  Where else could one turn?

During this first quarter of 2005, many new voices appeared, many bonds were forged.

In this time before any of the other long running series, one man took the first plunge into the kiddie pool, every day, day after day, providing a safe and encouraging and, yes FUN and funny spot to get acquainted, rub elbows with like minded folks, and generally look forward instead of back.

That man was, of course, Bill in Portland Maine.

Bill unlocked the secret to series blogging on daily kos: post every single day, be funny and topical, and participate in the comment thread.  How he has managed to do this continuously, with only occasional bathroom breaks, for at least 13 years, is another mystery of nature.  Importantly for our story today, a good cross section of registered users (kossacks, kogs, kossarians) recommend and comment daily at Cheers and Jeers

But I come not just to praise Billy, but to relate how his Cheers & Jeers became the standard of comparison for the attention given to a post/diary here at daily kos.

Unlike others who appear to be able to pay constant attention to happenings in the Wonderful World of marKos, I am a bit slow and inattentive, and back in those days, found myself constantly wondering about references to on site events both large and small that I had no knowledge of.

What to do?

My solution was to "simply" page scrape every single diary that was posted, enter them into a database on my computer, and the next day look up which diaries got the most attention, where attention = number of recommendations.

But that didn't always work as expected - there were some diaries that engendered a large number of comments but very few recommendations.  Despite my hope that some of these might be the result of high minded discussions among those who eschew the fame and fortune that follows the highly recommended author (heh), in fact these sort of posts/diaries were almost invariably train wrecks: they presented themselves in such a way as to engender a large amount of mostly negative comment without any redeeming features that would collect recommendations.

So how do I automatically find those?

Well, I needed to somehow combine the number of recommendations and the number of comments in such a way that train wrecks would show up.  Moreover, I needed a single number that would reflect that combination so I could use it to sort the posts and make a table useful for reviewing the previous days action.

After a number of false starts, I hit on the winning formula: I would use a reference standard.  

The logical standard at that time was clearly Cheers and Jeers.  It appeared daily, it garnered a fairly steady number of comments and recommendations, and everyone knew what it was.

I present here the actual code used initially, along with comments on its origin.  I should note in addition that "bharn" was at that time an email address of BiPM.


def timpact(nrec,ncom):
        # calculate impact relative to a standard diary
        # namely BiPM's Cheers and Jeers
        # Between Jan 1, 2005 and April 4 2005 there were
        # 53 new week day issues of C&J. The average number of
        # recommendations received was 79.71, stdev = 11.38
        # and the average number of comments added to each was
        # 412.69, stdev = 72.71.
        # Defining these levels of recommendations and
        # thread inducement equivalent to 1 bharns of impact
        # and giving equal weight to each measure, it is
        # possible to calculate the relative impact of any diary,
        # given the number of recommendations (nrec) and the
        # number of comments, ncom, as the absolute value of the
        # imaginary number (nrec/79.71, ncom/412.69 i),
        # divided by the square root of two.
        # 78.9273       412.1091        101.4545
        rimpact =  nrec / 78.9273
        cimpact =  ncom / 412.1091
        impact = complex(rimpact,cimpact)
        return abs(impact)/sq2

This function is no longer in use, superseded by an improved formula that counts the number of users that recommend and the number of users that comment, thus downrating a diary with a very few very active commenters.

Impact is the number I use every day in my series High Impact Posts.

So BiPM and Cheers and Jeers have served as a standard for comparison for others efforts at daily kos.

This does not stop with my arcane hijacking of their "numbers" for a reference standard, as BiPM and Cheers and Jeers are an integral part of the community which is Daily Kos.

Our support and backing is crucial to the continued existence of Cheers and Jeers, as best explained by Bill.

I hope you've enjoyed this perhaps more than slightly technical/geeky/nerdy story, and the you will support Bill and C&J with your hard cold cash, at the following locations.

Thank you Bill!

One time contribution: click here.

$5 monthly contribution: click here

$10 monthly contribution: click here

$20 monthly contribution: click here

To send a donation via snail mail, the address is: Bill Harnsberger, 16 Pitt Street, Portland, ME, 04103.

Originally posted to jotter on Tue Oct 01, 2013 at 01:24 PM PDT.

Also republished by J Town and Pink Clubhouse.

EMAIL TO A FRIEND X
Your Email has been sent.