Board Changes: 1-2 May 2003
The MTU (today, about 2 days 20 hours) is the average time that topics, both
currently spoiled or unspoiled, on the board (non-announcements) have remained
unspoiled. This is the primary indicator of turnaround. If this time gets
longer and longer, we know we have many topics that have lain around for
a long time that are still unspoiled.
The MTS (today, about 10 hours 20 minutes) is the average time that a now-spoiled
topic on the board remained unspoiled. If long-incorrupt topics are spoiled,
this increases the MTS time; if pre-spoiled topics are added, this lowers
the MTS time.
Note that currently MTS time is much less than the MTU time, but
this need not be so - in fact it's possible for the MTS to be greater than the MTU.
If the MTS approaches the MTU, this is a sign that topics are being spoiled
evenly throughout the entire list, both old and new.
Also, the longer a topic goes unspoiled beyond the MTS time, the more likely
it is to be pushed back in the list and only be spoiled as the result of
a BM challenge or a Featured Topic placement.
These statistics, like the number of posts or users, are being continuously
tracked by the Broken Memories board script (rapidly approaching 6,000 lines of elegant C++ code) so
advanced statistical reports can be generated for my own edification and bizarre amusement.
This number is also used in awarding Sun of Purity medals (the second-highest medal
on the site). Further, 10 points are awarded to the topic creator when
the topic reaches the List of Incorruptibles. (If the topic is knocked off
the list and reascends later into the rank of Incorruptibles, no second
award is given.)
Please note that the way we weight this currently only depreciates topics
with few votes. It does
not make topics with few votes tend to an "average" weighted rating, it makes
them look bad. Thus a topic with a couple of 1 votes and no others has a very, very low
weighted
rating indeed. This is done in the expectation that people will be searching
for topics that are voted the highest, not the lowest (we are sorting in
descending order, after all.) It also keeps consistent with the fact that
topics with zero votes are given, effectively, a zero rating and plummet to
the bottom of the list sorted by average vote. If experience shows me that this is not the
way the feature is being used, I will introduce a more traditional weighting scheme.
(for change is the only constant)