Amiga.org

Coffee House => Coffee House Boards => CH / General => Topic started by: Floid on April 08, 2004, 09:18:07 AM

Title: Proposal: New Metrics for Probabilistic Determination of Productivity in Online Discussion
Post by: Floid on April 08, 2004, 09:18:07 AM
Note that this is not a law, as in Murphy's or Godwin's, but a scale of measurement for an observable property.

Assumed that the productivity* of a discussion is halved for each response to a poster questioning his own involvement, we can derive a score by which to determine the noise of a channel (in this case, a thread), and from this predict our chances of communicating successfully through it.

Proposed are two units, the Amon and the Alemagna.  Amon are a simple count; thus a single response to "Why does my name keep popping up in this thread?" gives a thread a rating of 1 Amon.  Alemagna are further derived units, taken by dividing the initial productivity of a channel (which may, but does not have to, default to 1**) by half for each Amon it rates.

Thus:

    Alemagna = 1 / 2^Amon

So, a thread with an initially assumed productivity of 1 and carrying a rating of 1 Amon works out to a score of 0.5 Alemagna; a 2 Amon thread becomes 0.25 Alemagna, and so forth.  Correlation is unproven, but the Alemagna rating may be considered a rough probability of further discussion being noticed or taken seriously.

*'Productivity' is here defined as the percentage probability of new information being conveyed to the audience of the forum as a whole.

**Other factors may influence the initial productivity of a thread, ranging from anonymity on the part of the first poster to assertions of trolling.

---

[All in good fun, guys!]
Title: Re: Proposal: New Metrics for Probabilistic Determination of Productivity in Online Discussion
Post by: Karlos on April 08, 2004, 10:39:33 AM
I can see the productivity of this thread having an exponential decay in relation to the number of posts without the formal definition ;-)