[pmwiki-users] Improving PITS
Kathryn Andersen
kat_lists at katspace.homelinux.org
Wed Jul 8 21:08:54 CDT 2009
On Wed, Jul 08, 2009 at 09:37:07PM -0400, DaveG wrote:
> Petko Yotov wrote:
>> On Wednesday 08 July 2009 20:02:48 Kathryn Andersen wrote:
>> I don't want entries with a single vote "5" ordered before entries with
>> many votes like "55555 331". For me, votes "55444431X" are of higher
>> priority than one vote "5", even if "5444431X" has an average lower
>> than 4. That's why there is a lower limit.
> I'm not a mathematician, but perhaps there is some algorithm we can
> apply such that entries with a lot of votes get weighted differntly to
> those with few.
>
> Thus some thing with a spread of votes 55555 44444 33333 will get a 4.7
> average (rather than the 'obvious' 4), because of the number of votes.
> Whereas something with a single 5 gets a 3.5, because it doesn't yet
> have a 'critical mass' of support.
I like that idea, if we could come up with an algorithm which is clear
and fair. Hmmm.
What if the "average" were calculated such that, for entries with less
than N votes, it was calculated such that the "missing" votes were zero?
For example, if N is 4, then an entry with a single vote of "5" would
get calculated as if the vote was 5 0 0 0 instead. That would give a
weighted average of 1.25.
5 0 0 0 -> 1.25
5 5 0 0 -> 2.5
5 5 5 0 -> 4
5 5 5 5 -> 5
This would mean that having a critical mass of even low-priority would
boost the weighted average; for example 5 1 1 1 gives 2 rather than 1.25
in this setup.
This doesn't give entries with lots of votes more weight, but it does
give entries with very few votes less weight.
Kathryn Andersen
--
_--_|\ | Kathryn Andersen <http://www.katspace.com>
/ \ |
\_.--.*/ | GenFicCrit mailing list <http://www.katspace.com/gen_fic_crit/>
v |
------------| Melbourne -> Victoria -> Australia -> Southern Hemisphere
Maranatha! | -> Earth -> Sol -> Milky Way Galaxy -> Universe
More information about the pmwiki-users
mailing list