[extropy-chat] META: ExI List Quality & Future

[Hal Finney]

On 2/21/06, Adrian Tymes <wingcat at pacbell.net> wrote:
> How about this?  Anyone who wants to help moderate the list, can
> register for that Web board with the same email address as used
> to subscribe to this list.  Then, whenever they see something
> that in their opinion requires moderation, they log on to the
> board and give karma, Slashdot-style (although, not tracked
> long-term), with negative and positive karma cancelling out.

We had that "karma" concept on the message board a couple of years
ago.  It didn't seem to work too well.  Here is a posting I made back
on March 11, 2003, talking about the web board:

: As you read
: the messages you see not only a picture that the poster has chosen to
: associate with himself (a very big one for Lee Crocker!), but also the
: poster's "karma" score.  Anyone can vote positively or negatively for
: another person when reading their posts.  The postive and negative
: values are apparently totalled separately.
:
: The karma scores seem somewhat counter-intuitive.  It's not surprising
: to see Anders Sandburg at +235/-1 or Eliezer Yudkowski at +123/-1.
: Amara Graps seems to be the champion at +273/-4.  And apparently Harvey
: Newstrom hasn't gotten anybody mad at him, at +157/-0.
:
: But Spike at +4/-213?  That seems a little harsh.  Lee Corbin at +4/-305,
: Lee Crocker at +1/-52, John Clark at +5/-224?  These are some of the
: people I find most informative and stimulating.  Rafal Smigrodzki at
: +1/-296, Robert Bradbury at +17/-187?  I don't really get it.
:
: My own score is a mediocre +8/-46.  Pretty discouraging.  Lately I
: mostly use the web interface so these numbers are being rubbed in my
: face constantly.
:
: Hopefully the database is keeping track of who is voting for whom.
: It would be interesting to use the data in more sophisticated ways,
: such as looking for clustering effects.  Maybe all those negative karma
: points were applied by a relatively self-contained group.  It would be
: interesting to see more about how the scores have evolved.

Although I did not say it at the time, it looked to me like many of
these scores were applied based on political orientation.  People who
expressed conservative or libertarian views had strong negative scores,
while those with more left-leaning positions had a positive boost.
And the numbers were so disproportionately large, it seemed likely that`
it was a script which had hit the board repeatedly and somehow managed to
"stuff the ballot box" with multiple votes reflecting the perpetrators's
leftist political views.

Since then the karma scores have been cleared, so the evidence of what
happened is presumably lost.  But clearly any scheme for automated
"karma" scoring needs to protect against multiple voting.

Hal